Provided by: stress-ng_0.19.02-1_amd64 
NAME
stress-ng - stress "next generation", a tool to load and stress a computer system
SYNOPSIS
stress-ng [OPTION [ARG]] ...
DESCRIPTION
stress-ng will stress test a computer system in various selectable ways. It was designed to exercise
physical subsystems of a computer as well as operating system kernel interfaces. stress-ng also has a
wide range of CPU specific stress tests that exercise floating point, integer, bit manipulation, cache,
vector instructions and control flow.
stress-ng was originally intended to make a machine work hard and trip hardware issues such as thermal
overruns as well as operating system bugs that only occur when a system is being thrashed hard. Use
stress-ng with caution as some of the tests can make a system run hot on poorly designed hardware and
also can cause excessive system thrashing which may be difficult to stop.
stress-ng can also measure test throughput rates; this can be useful to observe performance changes
across different operating system releases or types of hardware. However, it has never been intended to
be used as a precise benchmark test suite, so do NOT use it in this manner.
Running stress-ng with root privileges will adjust out of memory settings on Linux systems to make the
stressors unkillable in low memory situations, so use this judiciously. With the appropriate privilege,
stress-ng can allow the ionice class and ionice levels to be adjusted, again, this should be used with
care.
One can specify the number of processes to invoke per type of stress test; specifying a zero value will
select the number of processors available as defined by sysconf(_SC_NPROCESSORS_CONF), if that can't be
determined then the number of online CPUs is used. If the value is less than zero then the number of
online CPUs is used. Specifying the number as a percentage will select the percentage of configured CPUs
(truncated down to nearest whole number).
OPTIONS
General stress-ng control options:
--abort
this option will force all running stressors to abort (terminate) if any other stressor terminates
prematurely because of a failure.
--aggressive
enables more file, cache and memory aggressive options. This may slow tests down, increase
latencies and reduce the number of bogo ops as well as changing the balance of user time vs system
time used depending on the type of stressor being used.
-a N, --all N, --parallel N
start N instances of all stressors in parallel. If N is less than zero, then the number of CPUs
online is used for the number of instances. If N is zero, then the number of configured CPUs in
the system is used.
-b N, --backoff N
wait N microseconds between the start of each stress worker process. This allows one to ramp up
the stress tests over time.
--buildinfo
report build information such as build date and time, compiler used, compilation and linker flags,
standard C version used.
--c-states
report CPU C-state residencies.
--change-cpu
this forces child processes of some stressors to change to a different CPU from the parent on
startup. Note that during the execution of the stressor the scheduler may choose move the parent
onto the same CPU as the child. The stressors affected by this option are client/server style
stressors, such as the network stressors (sock, sockmany, udp, etc) or context switching stressors
(switch, pipe, etc).
--class name
specify the class of stressors to run. Stressors are classified into one or more of the following
classes: compute, cpu, cpu-cache, device, fp, gpu, interrupt, integer, ipc, io, filesystem,
memory, network, os, pipe, scheduler, search, signal, sort, vector and vm. Some stressors fall
into just one class. For example the `get' stressor is just in the `os' class. Other stressors
fall into more than one class, for example, the `lsearch' stressor falls into the `cpu',
`cpu-cache', `memory' and `search' classes as it exercises all these four. Selecting a specific
class will run all the stressors that fall into that class only when run with the --sequential
option.
Specifying a name followed by an escaped question mark (for example --class vm\?) will print out
all the stressors in that specific class.
--config
print out the configuration used to build stress-ng.
-n, --dry-run
parse options, but do not run stress tests. A no-op.
--ftrace
enable kernel function call tracing (Linux only). This will use the kernel debugfs ftrace
mechanism to record all the kernel functions used on the system while stress-ng is running. This
is only as accurate as the kernel ftrace output, so there may be some variability on the data
reported.
-h, --help
show help.
--ignite-cpu
alter kernel controls to try and maximize the CPU. This requires root privilege to alter various
/sys interface controls. Currently this only works for Intel P-State enabled x86 systems on
Linux.
--interrupts
check for any system management interrupts or error interrupts that occur, for example thermal
overruns, machine check exceptions, etc. Note that the interrupts are accounted to all the
concurrently running stressors, so total count for all stressors is over accounted.
--ionice-class class
specify ionice class (only on Linux). Can be idle (default), besteffort, be, realtime, rt.
--ionice-level level
specify ionice level (only on Linux). For idle, 0 is the only possible option. For besteffort or
realtime values 0 (highest priority) to 7 (lowest priority). See ionice(1) for more details.
--iostat S
every S seconds show I/O statistics on the device that stores the stress-ng temporary files. This
is either the device of the current working directory or the --temp-path specified path. Currently
a Linux only option. The fields output are:
Column Heading Explanation
Inflight number of I/O requests that have been issued to the device driver but have not
yet completed
Rd K/s read rate in 1024 bytes per second
Wr K/s write rate in 1024 bytes per second
Rd/s reads per second
Wr/s writes per second
--job jobfile
run stressors using a jobfile. The jobfile is essentially a file containing stress-ng options
(without the leading --) with one option per line. Lines may have comments with comment text
proceeded by the # character. A simple example is as follows:
run sequential # run stressors sequentially
verbose # verbose output
metrics-brief # show metrics at end of run
timeout 60s # stop each stressor after 60 seconds
#
# vm stressor options:
#
vm 2 # 2 vm stressors
vm-bytes 128M # 128 MB available memory shared by 2 vm stressors
vm-keep # keep vm mapping
vm-populate # populate memory
#
# memcpy stressor options:
#
memcpy 5 # 5 memcpy stressors
The job file introduces the run command that specifies how to run the stressors:
run sequential - run stressors sequentially
run parallel - run stressors together in parallel
Note that `run parallel' is the default.
--keep-files
do not remove files and directories created by the stressors. This can be useful for debugging
purposes. Not generally recommended as it can fill up a file system.
-k, --keep-name
by default, stress-ng will attempt to change the name of the stress processes according to their
functionality; this option disables this and keeps the process names to be the name of the parent
process, that is, stress-ng.
-K, --klog-check
check the kernel log for kernel error and warning messages and report these as soon as they are
detected. Linux only and requires root capability to read the kernel log.
--ksm enable kernel samepage merging (Linux only). This is a memory-saving de-duplication feature for
merging anonymous (private) pages.
--log-brief
by default stress-ng will report the name of the program, the message type and the process id as a
prefix to all output. The --log-brief option will output messages without these fields to produce
a less verbose output.
--log-file filename
write messages to the specified log file, this does not replace the output from stdout, but
creates a logged copy of it in the specified file. Lines of output are automatically fsync'd to
the log file to try to reduce data loss.
--log-lockless
log messages use a lock to avoid intermingling of blocks of stressor messages, however this may
cause contention when emitting a high rate of logging messages in verbose mode with many stressors
are running, for example when testing CPU scaling with many processes on many CPUs. This option
disables log message locking.
--maximize
overrides the default stressor settings and instead sets these to the maximum settings allowed.
These defaults can always be overridden by the per stressor settings options if required.
--max-fd N
set the maximum limit on file descriptors (value or a % of system allowed maximum). By default,
stress-ng can use all the available file descriptors; this option sets the limit in the range from
10 up to the maximum limit of RLIMIT_NOFILE. One can use a % setting too, e.g. 50% is half the
maximum allowed file descriptors. Note that stress-ng will use about 5 of the available file
descriptors so take this into consideration when using this setting.
--mbind list
set strict NUMA memory allocation based on the list of NUMA nodes provided; page allocations will
come from the node with sufficient free memory closest to the specified node(s) where the
allocation takes place. This uses the Linux set_mempolicy(2) call using the MPOL_BIND mode. The
NUMA nodes to be used are specified by a comma separated list of node (0 to N-1). One can specify
a range of NUMA nodes using `-', for example: --mbind 0,2-3,6,7-11
-M, --metrics
output number of bogo operations in total performed by the stress processes. Note that these are
not a reliable metric of performance or throughput and have not been designed to be used for
benchmarking whatsoever. Some stressors have additional metrics that are more useful than bogo-
ops, and these are generally more useful for observing how a system behaves when under various
kinds of load.
The following columns of information are output:
Column Heading Explanation
bogo ops number of iterations of the stressor during the run. This is metric of
how much overall "work" has been achieved in bogo operations. Do not
use this as a reliable measure of throughput for benchmarking.
real time (secs) average wall clock duration (in seconds) of the stressor. This is the
total wall clock time of all the instances of that particular stressor
divided by the number of these stressors being run.
usr time (secs) total user time (in seconds) consumed running all the instances of the
stressor.
sys time (secs) total system time (in seconds) consumed running all the instances of
the stressor.
bogo ops/s (real time) total bogo operations per second based on wall clock run time. The
wall clock time reflects the apparent run time. The more processors
one has on a system the more the work load can be distributed onto
these and hence the wall clock time will reduce and the bogo ops rate
will increase. This is essentially the "apparent" bogo ops rate of
the system.
bogo ops/s (usr+sys time) total bogo operations per second based on cumulative user and system
time. This is the real bogo ops rate of the system taking into
consideration the actual time execution time of the stressor across
all the processors. Generally this will decrease as one adds more
concurrent stressors due to contention on cache, memory, execution
units, buses and I/O devices.
CPU used per instance (%) total percentage of CPU used divided by number of stressor instances.
100% is 1 full CPU. Some stressors run multiple threads so it is
possible to have a figure greater than 100%.
RSS Max (KB) resident set size (RSS), the portion of memory (measured in Kilobytes)
occupied by a process in main memory.
--metrics-brief
show shorter list of stressor metrics (no CPU used per instance).
--minimize
overrides the default stressor settings and instead sets these to the minimum settings allowed.
These defaults can always be overridden by the per stressor settings options if required.
--no-madvise
from version 0.02.26 stress-ng automatically calls madvise(2) with random advise options before
each mmap and munmap to stress the vm subsystem a little harder. The --no-advise option turns this
default off.
--no-oom-adjust
disable any form of out-of-memory score adjustments, keep the system defaults. Normally stress-ng
will adjust the out-of-memory scores on stressors to try to create more memory pressure. This
option disables the adjustments.
--no-rand-seed
Do not seed the stress-ng pseudo-random number generator with a quasi random start seed, but
instead seed it with constant values. This forces tests to run each time using the same start
conditions which can be useful when one requires reproducible stress tests.
--oom-avoid
Attempt to avoid out-of-memory conditions that can lead to the Out-of-Memory (OOM) killer
terminating stressors. This checks for low memory scenarios and swapping before making memory
allocations and hence adds some overhead to the stressors and will slow down stressor allocation
speeds.
--oom-avoid-bytes N
Specify a low memory threshold to avoid making any further memory allocations. The parameter can
be specified as an absolute number of bytes (e.g. 2M for 2 MB) or a percentage of the current free
memory, e.g. 5% (the default is 2.5%). This option implicitly enables --oom-avoid. The option
allows the system to have enough free memory to try to avoid the out-of-memory killer terminating
processes.
--oomable
Do not respawn a stressor if it gets killed by the Out-of-Memory (OOM) killer. The default
behaviour is to restart a new instance of a stressor if the kernel OOM killer terminates the
process. This option disables this default behaviour.
--page-in
touch allocated pages that are not in core, forcing them to be paged back in. This is a useful
option to force all the allocated pages to be paged in when using the bigheap, mmap and vm
stressors. It will severely degrade performance when the memory in the system is less than the
allocated buffer sizes. This uses mincore(2) to determine the pages that are not in core and
hence need touching to page them back in.
--pathological
enable stressors that are known to hang systems. Some stressors can rapidly consume resources that
may hang a system, or perform actions that can lock a system up or cause it to reboot. These
stressors are not enabled by default, this option enables them, but you probably don't want to do
this. You have been warned. This option applies to the stressors: bad-ioctl, bind-mount,
cpu-online, mlockmany, oom-pipe, smi, sysinval and watchdog.
--pause T
pause T seconds between each stressor. This is useful for allowing systems to cool down between
each stressor invocation. By default this option is disabled.
--perf measure processor and system activity using perf events. Linux only and caveat emptor, according
to perf_event_open(2): "Always double-check your results! Various generalized events have had
wrong values". Note that with Linux 4.7 one needs to have CAP_SYS_ADMIN capabilities for this
option to work, or adjust /proc/sys/kernel/perf_event_paranoid to below 2 to use this without
CAP_SYS_ADMIN.
--permute N
run all permutations of the selected stressors with N instances of the permutated stressors per
run. If N is less than zero, then the number of CPUs online is used for the number of instances.
If N is zero, then the number of configured CPUs in the system is used. This will perform
multiple runs with all the permutations of the stressors up to 2↑20 permumtations. Use this in
conjunction with the --with or --class option to specify the stressors to permute.
--progress
display the run progress when running stressors with the --sequential option.
-q, --quiet
do not show any output.
-r N, --random N
start N random stress workers. If N is 0, then the number of configured processors is used for N.
--rapl Report the Running Average Power Limit (RAPL) energy measurements of stressor instances. Currently
Linux and x86 only, requires root access rights to read RAPL kernel interfaces. Note that the RAPL
domains supported may vary between devices.
--raplstat S
every S seconds show RAPL energy measurements. Currently Linux and x86 only, requires root access
rights to read RAPL kernel interfaces.
--sched scheduler
select the named scheduler (only on Linux). To see the list of available schedulers use: stress-ng
--sched which
--sched-prio prio
select the scheduler priority level (only on Linux). If the scheduler does not support this then
the default priority level of 0 is chosen.
--sched-period period
select the period parameter for deadline scheduler (only on Linux). Default value is 0 (in
nanoseconds).
--sched-runtime runtime
select the runtime parameter for deadline scheduler (only on Linux). Default value is 99999 (in
nanoseconds).
--sched-deadline deadline
select the deadline parameter for deadline scheduler (only on Linux). Default value is 100000 (in
nanoseconds).
--sched-reclaim
use cpu bandwidth reclaim feature for deadline scheduler (only on Linux).
--seed N
set the random number generate seed with a 64 bit value. Allows stressors to use the same random
number generator sequences on each invocation.
--settings
show the various option settings.
--sequential N
sequentially run all the stressors one by one for a default of 60 seconds. The number of instances
of each of the individual stressors to be started is N. If N is less than zero, then the number
of CPUs online is used for the number of instances. If N is zero, then the number of CPUs in the
system is used. Use the --timeout option to specify the duration to run each stressor.
--skip-silent
silence messages that report that a stressor has been skipped because it requires features not
supported by the system, such as unimplemented system calls, missing resources or processor
specific features.
--smart
scan the block devices for changes S.M.A.R.T. statistics (Linux only). This requires root
privileges to read the Self-Monitoring, Analysis and Reporting Technology data from all block
devies and will report any changes in the statistics. One caveat is that device manufacturers
provide different sets of data, the exact meaning of the data can be vague and the data may be
inaccurate.
--sn use scientific notation (e.g. 2.412e+01) for metrics.
--status N
report every N seconds the number of running, exiting, reaped and failed stressors, number of
stressors that received SIGARLM termination signal as well as the current run duration.
--stderr
write messages to stderr. With version 0.15.08 output is written to stdout, previously due to a
historical oversight output went to stderr. This option allows one to revert to the pre-0.15.08
behaviour.
--stdout
all output goes to stdout. This is the new default for version 0.15.08. Use the --stderr option
for the original behaviour.
--stressor-time
(requires -v flag) log as debug the start and finish run times of each stressor instance, logged
in the format: stressor [start|finish] HR:MN:SS.HS YYYY:MM:DD where HR is hours, MN is minutes, SS
is seconds, HS is hundredths of seconds, YYYY is years, MM is months, DD is day of the month. If
localtime(2) is not supported the time is logged as seconds past the Epoch.
--stressors
output the names of the available stressors.
--sync-start
synchromize start, wait for stressors to be created and start all stressors once they are all in a
ready to run state.
--syslog
log output (except for verbose -v messages) to the syslog.
--taskset list
set CPU affinity based on the list of CPUs provided; stress-ng is bound to just use these CPUs
(for systems that provide sched_setaffinity()). The CPUs to be used are specified by a comma
separated list of CPU (0 to N-1). One can specify a range of CPUs using `-', for example:
--taskset 0,2-3,6,7-11 or the following keywords:
Keyword Description
even even numbered CPUs
odd odd numbered CPUs
all all CPUs
random random selection of CPUs
packageN CPUs in package N as specified by
/sys/devices/system/cpu/cpu*/topology/package_cpus_list
clusterN CPUs in package N as specified by
/sys/devices/system/cpu/cpu*/topology/cluster_cpus_list
dieN CPUs in package N as specified by /sys/devices/system/cpu/cpu*/topology/die_cpus_list
coreN CPUs in package N as specified by /sys/devices/system/cpu/cpu*/topology/core_cpus_list
--taskset-random
randomly change stressor CPU affinity at five times the clock tick rate (if defined) or at 400Hz
while waiting for stressors to complete.
--temp-path path
specify a path for stress-ng temporary directories and temporary files; the default path is the
current working directory. This path must have read and write access for the stress-ng stress
processes.
--thermalstat S
every S seconds show CPU and thermal load statistics. This option shows average CPU frequency in
GHz (average of online-CPUs), the minimum CPU frequency, the maximum CPU frequency, load averages
(1 minute, 5 minute and 15 minutes) and available thermal zone temperatures in degrees Centigrade.
--thrash
This can only be used when running on Linux and with root privilege. This option starts a
background thrasher process that works through all the processes on a system and tries to page-in
as many pages in the processes as possible. It also periodically drops the page cache, frees
reclaimable slab objects and pagecache as well as assign pages to different NUMA nodes. This will
cause considerable amount of thrashing of swap on an over-committed system.
-t N, --timeout T
run each stress test for at least T seconds. One can also specify the units of time in seconds,
minutes, hours, days or years with the suffix s, m, h, d or y. Each stressor will be sent a
SIGALRM signal at the timeout time, however if the stress test is swapped out, in an
uninterruptible system call or performing clean up (such as removing hundreds of test file) it may
take a while to finally terminate. A 0 timeout will run stress-ng for ever with no timeout. The
default timeout is 24 hours.
--times
show the cumulative user and system times of all the child processes at the end of the stress run.
The percentage of utilisation of available CPU time is also calculated from the number of on-line
CPUs in the system.
--timestamp
add a timestamp in hours, minutes, seconds and hundredths of a second to the log output.
--timer-slack N
adjust the per process timer slack to N nanoseconds (Linux only). Increasing the timer slack
allows the kernel to coalesce timer events by adding some fuzziness to timer expiration times and
hence reduce wakeups. Conversely, decreasing the timer slack will increase wakeups. A value of 0
for the timer-slack will set the system default of 50,000 nanoseconds.
--tz collect temperatures from the available thermal zones on the machine (Linux only). Some devices
may have one or more thermal zones, where as others may have none.
-v, --verbose
show all debug, warnings and normal information output.
--verify
verify results when a test is run. This is not available on all tests. This will sanity check the
computations or memory contents from a test run and report to stderr any unexpected failures.
--verifiable
print the names of stressors that can be verified with the --verify option.
-V, --version
show version of stress-ng, version of toolchain used to build stress-ng and system information.
--vmstat S
every S seconds show statistics about processes, memory, paging, block I/O, interrupts, context
switches, disks and cpu activity. The output is similar that to the output from the vmstat(8)
utility. Not fully supported on various UNIX systems.
--vmstat-units [ k | m | g | t | p | e ]
specify vmstat memory units in terms of kilobytes (k), megabytes (m), gigabytes (g), terabytes
(t), petabytes (p) or exabytes (e). Default is in kilobytes.
-w, --with list
specify stressors to run when using the --all, --seq or --permute options. For example to run 5
instances of the cpu, hash, nop and vm stressors one after another (sequentially) for 1 minute per
stressor use:
stress-ng --seq 5 --with cpu,hash,nop,vm --timeout 1m
-x, --exclude list
specify a list of one or more stressors to exclude (that is, do not run them). This is useful to
exclude specific stressors when one selects many stressors to run using the --class option,
--sequential, --all and --random options. Example, run the cpu class stressors concurrently and
exclude the numa and search stressors:
stress-ng --class cpu --all 1 -x numa,bsearch,hsearch,lsearch
-Y, --yaml filename
output gathered statistics to a YAML formatted file named `filename'.
Stressor specific options:
Access stressor
--access N
start N workers that work through various settings of file mode bits (read, write, execute)
for the file owner and checks if the user permissions of the file using access(2) and
faccessat(2) are sane.
--access-ops N
stop access workers after N bogo access sanity checks.
POSIX Access Control List Stressor
--acl N
start N workers that exercise permutations of ACL access permission settings on user, group
and other tags.
--acl-rand
randomize (by shuffling) the order of the ACL access permissions before exercising ACLs.
--acl-ops N
stop acl workers after N bogo acl settings have been set.
Affinity stressor
--affinity N
start N workers that run 16 processes that rapidly change CPU affinity (for systems that
provide sched_setaffinity()). Rapidly switching CPU affinity can contribute to poor cache
behaviour and high context switch rate.
--affinity-delay N
delay for N nanoseconds before changing affinity to the next CPU. The delay will spin on CPU
scheduling yield operations for N nanoseconds before the process is moved to another CPU. The
default is 0 nanosconds.
--affinity-ops N
stop affinity workers after N bogo affinity operations.
--affinity-pin
pin all the 16 per stressor processes to a CPU. All 16 processes follow the CPU chosen by the
main parent stressor, forcing heavy per CPU loading.
--affinity-rand
switch CPU affinity randomly rather than the default of sequentially.
--affinity-sleep N
sleep for N nanoseconds before changing affinity to the next CPU.
Kernel crypto AF_ALG API stressor
--af-alg N
start N workers that exercise the AF_ALG socket domain by hashing and encrypting various
sized random messages. This exercises the available hashes, ciphers, rng and aead crypto
engines in the Linux kernel.
--af-alg-dump
dump the internal list representing cryptographic algorithms parsed from the /proc/crypto
file to standard output (stdout).
--af-alg-ops N
stop af-alg workers after N AF_ALG messages are hashed.
Asynchronous I/O stressor (POSIX AIO)
--aio N
start N workers that issue multiple small asynchronous I/O writes and reads on a relatively
small temporary file using the POSIX aio interface. This will just hit the file system cache
and soak up a lot of user and kernel time in issuing and handling I/O requests. By default,
each worker process will handle 16 concurrent I/O requests.
--aio-ops N
stop POSIX asynchronous I/O workers after N bogo asynchronous I/O requests.
--aio-requests N
specify the number of POSIX asynchronous I/O requests each worker should issue, the default
is 16; 1 to 4096 are allowed.
Asynchronous I/O stressor (Linux AIO)
--aiol N
start N workers that issue multiple 4 K random asynchronous I/O writes using the Linux aio
system calls io_setup(2), io_submit(2), io_getevents(2) and io_destroy(2). By default, each
worker process will handle 16 concurrent I/O requests.
--aiol-ops N
stop Linux asynchronous I/O workers after N bogo asynchronous I/O requests.
--aiol-requests N
specify the number of Linux asynchronous I/O requests each worker should issue, the default
is 16; 1 to 4096 are allowed.
Alarm stressor
--alarm N
start N workers that exercise alarm(2) with MAXINT, 0 and random alarm and sleep delays that
get prematurely interrupted. Before each alarm is scheduled any previous pending alarms are
cancelled with zero second alarm calls.
--alarm-ops N
stop after N alarm bogo operations.
AppArmor stressor
--apparmor N
start N workers that exercise various parts of the AppArmor interface. Currently one needs
root permission to run this particular test. Only available on Linux systems with AppArmor
support and requires the CAP_MAC_ADMIN capability.
--apparmor-ops
stop the AppArmor workers after N bogo operations.
Atomic stressor
--atomic N
start N workers that exercise various GCC __atomic_*() built in operations on 8, 16, 32 and
64 bit integers that are shared among the N workers. This stressor is only available for
builds using GCC 4.7.4 or higher. The stressor forces many front end cache stalls and cache
references. Note that 32 bit systems do not currently exercise 64 bit integers.
--atomic-ops N
stop the atomic workers after N bogo atomic operations.
Bad alternative stack stressor
--bad-altstack N
start N workers that create broken alternative signal stacks for SIGSEGV and SIGBUS handling
that in turn create secondary SIGSEGV/SIGBUS errors. A variety of randomly selected
nefarious methods are used to create the stacks:
• Unmapping the alternative signal stack, before triggering the signal handling.
• Changing the alternative signal stack to just being read only, write only, execute only.
• Using a NULL alternative signal stack.
• Using the signal handler object as the alternative signal stack.
• Unmapping the alternative signal stack during execution of the signal handler.
• Using a read-only text segment for the alternative signal stack.
• Using an undersized alternative signal stack.
• Using the VDSO as an alternative signal stack.
• Using an alternative stack mapped onto /dev/zero.
• Using an alternative stack mapped to a zero sized temporary file to generate a SIGBUS
error.
--bad-altstack-ops N
stop the bad alternative stack stressors after N SIGSEGV bogo operations.
Bad ioctl stressor
--bad-ioctl N
start N workers that perform a range of illegal bad read ioctls (using _IOR) and write ioctls
(using _IOR with PROT_NONE mapped pages) across the device drivers. This exercises page size,
64 bit, 32 bit, 16 bit and 8 bit reads as well as NULL addresses, non-readable pages and
PROT_NONE mapped pages. Currently only for Linux and requires the --pathological option.
--bad-ioctl-method [ inc | random | random-inc | stride ]
select the method of changing the ioctl command (number, type) tuple per iteration, the
default is random-inc. Available bad-ioctl methods are described as follows:
Method Description
inc increment ioctl command by 1
random use a random ioctl command
random-inc increment ioctl command by a random value
random-stride increment ioctl command number by 1 and decrement command type by 3
--bad-ioctl-ops N
stop the bad ioctl stressors after N bogo ioctl operations.
Bessel functions
--besselmath N
start N workers that exercise various Bessel functions. Results are sanity checked to ensure
no variation occurs after each round of 10000 computations.
--besselmath-ops N
stop after N bessel bogo-operation loops.
--besselmath-method method
specify a Bessel function to exercise. Available bessel stress methods are described as
follows:
Method Description
all iterate over all the below Bessel functions methods
j0 double precision Bessel function of the first kind of order 0
j1 double precision Bessel function of the first kind of order 1
jn double precision Bessel function of the first kind of order n (where n = 5 for this
test)
j0f float precision Bessel function of the first kind of order 0
j1f float precision Bessel function of the first kind of order 1
jnf float precision Bessel function of the first kind of order n (where n = 5 for this
test)
j0l long double precision Bessel function of the first kind of order 0
j1l long double precision Bessel function of the first kind of order 1
jnl long double precision Bessel function of the first kind of order n (where n (\q 5 for
this test)
y0 double precision Bessel function of the second kind of order 0
y1 youble precision Bessel function of the second kind of order 1
yn double precision Bessel function of the second kind of order n (where n = 5 for this
test)
y0f float precision Bessel function of the second kind of order 0
y1f float precision Bessel function of the second kind of order 1
ynf float precision Bessel function of the second kind of order n (where n = 5 for this
test)
y0l long double precision Bessel function of the second kind of order 0
y1l long double precision Bessel function of the second kind of order 1
ynl long double precision Bessel function of the second kind of order n (where n = 5 for
this test)
Big heap stressor
-B N, --bigheap N
start N workers that grow their heaps by reallocating memory. If the out of memory killer
(OOM) on Linux kills the worker or the allocation fails then the allocating process starts
all over again. Note that the OOM adjustment for the worker is set so that the OOM killer
will treat these workers as the first candidate processes to kill.
--bigheap-bytes N
maximum heap growth as N bytes per bigheap worker. One can specify the size as % of total
available memory or in units of Bytes, KBytes, MBytes and GBytes using the suffix b, k, m or
g.
--bigheap-growth N
specify amount of memory to grow heap by per iteration. Size can be from 4 K to 64 MB.
Default is 64 K.
--bigheap-mlock
attempt to mlock future allocated pages into memory causing more memory pressure. If
mlock(MCL_FUTURE) is implemented then this will stop newly allocated pages from being swapped
out.
--bigheap-ops N
stop the big heap workers after N bogo allocation operations are completed.
Binderfs stressor
--binderfs N
start N workers that mount, exercise and unmount binderfs. The binder control device is
exercised with 256 sequential BINDER_CTL_ADD ioctl calls per loop.
--binderfs-ops N
stop after N binderfs cycles.
Bind mount stressor
--bind-mount N
start N workers that repeatedly bind mount / to / inside a user namespace. This can consume
resources rapidly, forcing out of memory situations. Do not use this stressor unless you want
to risk hanging your machine.
--bind-mount-ops N
stop after N bind mount bogo operations.
Bitonic sort stressor
--bitonicsort N
start N workers that sort 32 bit integers using bitonic sort.
--bitonicsort-ops N
stop bitonic sort stress workers after N bogo bitonic sorts.
--bitonicsort-size N
specify number of 32 bit integers to sort, default is 262144 (256 × 1024).
Bit Manipulation Operations
--bitops N
start N workers that perform various calculations using 32 bit integer manipulation
operations. Many of these are derived from the Standford "Bit Twiddling Hacks" (Sean Eron
Anderson) and Hacker's Delight (Henry S. Warren, Jr.)
--bitops-ops N
stop after N bitop operations.
--bitops-method method
specify bitops stress method. By default, all the bitops stress methods are exercised
sequentially, however one can specify just one method to be used if required. Available
bitops stress methods are described as follows:
Method Description
all iterate over all the below bitops stress methods.
abs calculate the absolute value of a 32 bit signed integer. Computed two different
ways, one with masking and addition, one with masking and subtraction.
countbits count the number of bits set to 1 in a 32 bit unsigned integer. Computed five
different ways, two with bit counting, one with 64 bit multiplication, one with
parallelised masking and shifting, one using a builtin popcount function.
clz count the number of leading zero bits in a 32 bit unsigned integer. Computed four
different ways, one with bit counting, one with log2 shifting, one using a builtin
popcount function, one using a builtin clz function,
ctz count the number of trailing zero bits in a 32 bit unsigned integer. Computed five
different ways, one with bit cointing, one using masking and shifting, one using
the masking and ternary operator Gaudet method, one using a builtin clz function,
one using a builtin popcount function.
cmp compare two 32 bit unsigned integers with comparison results of -1, 0, 1 for less
than, equal or greater than. Computed 3 ways, one using simple comparisons, two
using comparisons and subtraction.
log2 calculate log base 2 of a 32 bit unsigned integer. Computed four ways, one using
bit counting, one using masking and shifting and branching, one using masking and
shifting with no branching, one using shifting and bitwise or'ing.
max find the maximum of two 32 bit unsigned integers without a temporary variable.
Computed two ways, one using xor'ing and masking, one using the ternary operator.
min find the minimum of two 32 bit unsigned integers without a temporary variable.
Computed two ways, one using xor'ing and masking, one using the ternary operator.
parity compute the parity of a 32 bit unsigned integer. Computed five ways, two using bit
counting, one using multiplication, shifting and xor'ing, one using shifting and
xor'ing, one using a builtin parity function.
pwr2 determine if a 32 bit unsigned integer is a power of 2. Computed using bit
counting and with mask of value − 1.
rnddnpwr2 find the nearest power of 2 of a 32 bit unsigned integer, rounded down. Computed
three ways, one using bit counting and shifting, one using shifting and or'ing,
one using a shift and the builtin ctz function.
rnduppwr2 find the nearest power of 2 of a 32 bit unsigned integer, rounded up. Computed
three ways, one using bit counting and shifting, one using shifting and or'ing,
one using a shift and the builtin ctz function.
reverse reverse the bits of a 32 bit unsigned integer. Computed six ways, one using bit
shifting loop, one using bit shitting and mask loop, one using parallelised
masking and shifting, one using 64 bit multiplication, one using 32 bit
multiplication, one using the builtin bitreverse32 function.
sign calculate the sign of a 32 bit signed integer. Computed two ways, one using a
branchless less than zero operator, one using sign bit shifting and negation.
swap swap two 32 bit signed integers without a temporary variable. Computed two ways,
one using subtraction and addition, one using xor'ing.
zerobyte determine of a 32 bit unsigned integer contains one or more zero bytes. Computed
two ways, one with per byte zero checking, one using bit masking.
Branch stressor
--branch N
start N workers that randomly branch to 1024 randomly selected locations and hence exercise
the CPU branch prediction logic.
--branch-ops N
stop the branch stressors after N × 1024 branches
Brk stressor
--brk N
start N workers that grow the data segment by one page at a time using multiple brk(2) calls.
Each successfully allocated new page is touched to ensure it is resident in memory. If an
out of memory condition occurs then the test will reset the data segment to the point before
it started and repeat the data segment resizing over again. The process adjusts the out of
memory setting so that it may be killed by the out of memory (OOM) killer before other
processes. If it is killed by the OOM killer then it will be automatically re-started by a
monitoring parent process.
--brk-bytes N
maximum brk growth as N bytes per brk worker. One can specify the size as % of total
available memory or in units of Bytes, KBytes, MBytes and GBytes using the suffix b, k, m or
g.
--brk-mlock
attempt to mlock future brk pages into memory causing more memory pressure. If
mlock(MCL_FUTURE) is implemented then this will stop new brk pages from being swapped out.
--brk-notouch
do not touch each newly allocated data segment page. This disables the default of touching
each newly allocated page and hence avoids the kernel from necessarily backing the page with
physical memory.
--brk-ops N
stop the brk workers after N bogo brk operations.
Binary search stressor
--bsearch N
start N workers that binary search a sorted array of 32 bit integers using bsearch(3). By
default, there are 65536 elements in the array. This is a useful method to exercise random
access of memory and processor cache.
--bsearch-method [ bsearch-libc | bsearch-nonlibc | ternary ]
select either the libc implementation of bsearch or a slightly optimized non-libc
implementation of bsearch or a 3-day ternary search. The default is the libc implementation
if it exists, otherwise the non-libc version.
--bsearch-ops N
stop the bsearch worker after N bogo bsearch operations are completed.
--bsearch-size N
specify the size (number of 32 bit integers) in the array to bsearch. Size can be from 1 K to
4 M.
bubblesort stressor
--bubblesort N
start N workers that sort 32 bit integers using bubblesort.
--bubblesort-method [ bubblesort-fast | bubblesort-naive ]
select either a standard optimized implementation of bubblesort or a naive unoptimized
implementation of bubblesort. The default is the standard optimized version.
--bubblesort-ops N
stop bubblesort stress workers after N bogo bubblesorts.
--bubblesort-size N
specify number of 32 bit integers to sort, default is 16384.
Cache stressor
-C N, --cache N
start N workers that perform random wide spread memory read and writes to thrash the CPU
cache. The code does not intelligently determine the CPU cache configuration and so it may
be sub-optimal in producing hit-miss read/write activity for some processors. Note: to
exercise cache misses it is recommended to instead use the matrix-3d stressor using the
--matrix-3d-zyx option.
--cache-cldemote
cache line demote (x86 only). This is a no-op for non-x86 architectures and older x86
processors that do not support this feature.
--cache-clflushopt
use optimized cache line flush (x86 only). This is a no-op for non-x86 architectures and
older x86 processors that do not support this feature.
--cache-clwb
cache line writeback (x86 only). This is a no-op for non-x86 architectures and older x86
processors that do not support this feature.
--cache-enable-all
where appropriate exercise the cache using cldemote, clflushopt, fence, flush, sfence and
prefetch.
--cache-fence
force write serialization on each store operation (x86 only). This is a no-op for non-x86
architectures.
--cache-flush
force flush cache on each store operation (x86 only). This is a no-op for non-x86
architectures.
--cache-level N
specify level of cache to exercise (1 = L1 cache, 2 = L2 cache, 3 = L3/LLC cache (the
default)). If the cache hierarchy cannot be determined, built-in defaults will apply.
--cache-no-affinity
do not change processor affinity when --cache is in effect.
--cache-ops N
stop cache thrash workers after N bogo cache thrash operations.
--cache-permute
permute the selected cache flags when exercising cache for an improved mix of cache
operations. Recommend also using --cache-enable-all with this option.
--cache-prefetch
force read prefetch on read address on architectures that support prefetching.
--cache-prefetchw
force a redundant write prefetch on architectures that support write prefetching.
--cache-sfence
force write serialization on each store operation using the sfence instruction (x86 only).
This is a no-op for non-x86 architectures.
--cache-size N
override the default cache size setting to N bytes. One can specify the in units of Bytes,
KBytes, MBytes and GBytes using the suffix b, k, m or g.
--cache-ways N
specify the number of cache ways to exercise. This allows a subset of the overall cache size
to be exercised.
Cache hammering stessor
--cachehammer N
start N workers that exercise the cache with a randomized mix of memory read/writes and where
possible cache prefetches and cache flushes to random addresses in three memory mapped
regions, one of which is shared among all the cache stressors and is the size of the L3
cache, one is local to each instance and is 4 times the L3 cache size and one is a single
page mapping that cannot be read or written.
--cachehammer-numa
periodically assign exercised pages to randomly selected NUMA nodes. This is disabled for
systems that do not support NUMA.
--cachehammer-ops N
stop after N cache hammer operations.
Cache line stressor
--cacheline N
start N workers that exercise reading and writing individual bytes in a shared buffer that is
the size of a cache line. Each stressor has 2 running processes that exercise just two bytes
that are next to each other. The intent is to try and trigger cacheline corruption, stalls
and misses with shared memory accesses. For an N byte sized cacheline, it is recommended to
run N / 2 stressor instances.
--cacheline-affinity
frequently change CPU affinity, spread cacheline processes evenly across all online CPUs to
try and maximize lower-level cache activity. Attempts to keep adjacent cachelines being
exercised by adjacent CPUs.
--cacheline-method method
specify a cacheline stress method. By default, all the stress methods are exercised
sequentially, however one can specify just one method to be used if required. Available
cacheline stress methods are described as follows:
Method Description
all iterate over all the below cpu stress methods.
adjacent increment a specific byte in a cacheline and read the adjacent byte, check for
corruption every 7 increments.
atomicinc atomically increment a specific byte in a cacheline and check for corruption every
7 increments.
bits write and read back shifted bit patterns into specific byte in a cacheline and
check for corruption.
copy copy an adjacent byte to a specific byte in a cacheline.
inc increment and read back a specific byte in a cacheline and check for corruption
every 7 increments.
mix perform a mix of increment, left and right rotates a specific byte in a cacheline
and check for corruption.
rdfwd64 increment a specific byte in a cacheline and then read in forward direction an
entire cacheline using 64 bit reads.
rdints increment a specific byte in a cacheline and then read data at that byte location
in naturally aligned locations integer values of size 8, 16, 32, 64 and 128 bits.
rdrev64 increment a specific byte in a cacheline and then read in reverse direction an
entire cacheline using 64 bit reads.
rdwr read and write the same 8 bit value into a specific byte in a cacheline and check
for corruption.
--cacheline-ops N
stop cacheline workers after N loops of the byte exercising in a cacheline.
Process capabilities stressor
--cap N
start N workers that read per process capabilities via calls to capget(2) (Linux only).
--cap-ops N
stop after N cap bogo operations.
Cgroup stressor
--cgroup N
start N workers that mount a cgroup, move a child to the cgroup, read, write and remove the
child from the cgroup and umount the cgroup per bogo-op iteration. This uses cgroup v2 and
is only available for Linux systems.
--cgroup-ops N
stop after N cgroup bogo operations.
Chattr stressor
--chattr N
start N workers that attempt to exercise file attributes via the EXT2_IOC_SETFLAGS ioctl.
This is intended to be intentionally racy and exercise a range of chattr attributes by
enabling and disabling them on a file shared amongst the N chattr stressor processes. (Linux
only).
--chattr-ops N
stop after N chattr bogo operations.
Chdir stressor
--chdir N
start N workers that change directory between directories using chdir(2).
--chdir-dirs N
exercise chdir on N directories. The default is 8192 directories, this allows 64 to 65536
directories to be used instead.
--chdir-ops N
stop after N chdir bogo operations.
Chmod stressor
--chmod N
start N workers that change the file mode bits via chmod(2) and fchmod(2) on the same file.
The greater the value for N then the more contention on the single file. The stressor will
work through all the combination of mode bits.
--chmod-ops N
stop after N chmod bogo operations.
Chown stressor
--chown N
start N workers that exercise chown(2) on the same file. The greater the value for N then the
more contention on the single file.
--chown-ops N
stop the chown workers after N bogo chown(2) operations.
Chroot stressor
--chroot N
start N workers that exercise chroot(2) on various valid and invalid chroot paths. Only
available on Linux systems and requires the CAP_SYS_ADMIN capability.
--chroot-ops N
stop the chroot workers after N bogo chroot(2) operations.
Clock stressor
--clock N
start N workers exercising clocks and POSIX timers. For all known clock types this will
exercise clock_getres(2), clock_gettime(2) and clock_nanosleep(2). For all known timers it
will create a random duration timer and busy poll this until it expires. This stressor will
cause frequent context switching.
--clock-ops N
stop clock stress workers after N bogo operations.
Clone stressor
--clone N
start N workers that create clones (via the clone(2) and clone3() system calls). This will
rapidly try to create a default of 8192 clones that immediately die and wait in a zombie
state until they are reaped. Once the maximum number of clones is reached (or clone fails
because one has reached the maximum allowed) the oldest clone thread is reaped and a new
clone is then created in a first-in first-out manner, and then repeated. A random clone flag
is selected for each clone to try to exercise different clone operations. The clone stressor
is a Linux only option.
--clone-max N
try to create as many as N clone threads. This may not be reached if the system limit is less
than N.
--clone-ops N
stop clone stress workers after N bogo clone operations.
Close stressor
--close N
start N workers that try to force race conditions on closing opened file descriptors. These
file descriptors have been opened in various ways to try and exercise different kernel close
handlers.
--close-ops N
stop close workers after N bogo close operations.
Swapcontext stressor
--context N
start N workers that run three threads that use swapcontext(3) to implement the thread-to-
thread context switching. This exercises rapid process context saving and restoring and is
bandwidth limited by register and memory save and restore rates.
--context-ops N
stop context workers after N bogo context switches. In this stressor, 1 bogo op is
equivalent to 1000 swapcontext calls.
Copy file stressor
--copy-file N
start N stressors that copy a file using the Linux copy_file_range(2) system call. 128 KB
chunks of data are copied from random locations from one file to random locations to a
destination file. By default, the files are 256 MB in size. Data is sync'd to the filesystem
after each copy_file_range(2) call.
--copy-file-bytes N
copy file size, the default is 256 MB. One can specify the size as % of free space on the
file system or in units of Bytes, KBytes, MBytes and GBytes using the suffix b, k, m or g.
--copy-file-ops N
stop after N copy_file_range() calls.
CPU stressor
-c N, --cpu N
start N workers exercising the CPU by sequentially working through all the different CPU
stress methods. Instead of exercising all the CPU stress methods, one can specify a specific
CPU stress method with the --cpu-method option.
-l P, --cpu-load P
load CPU with P percent loading for the CPU stress workers. 0 is effectively a sleep (no
load) and 100 is full loading. The loading loop is broken into compute time (load%) and
sleep time (100% - load%). Accuracy depends on the overall load of the processor and the
responsiveness of the scheduler, so the actual load may be different from the desired load.
Note that the number of bogo CPU operations may not be linearly scaled with the load as some
systems employ CPU frequency scaling and so heavier loads produce an increased CPU frequency
and greater CPU bogo operations.
Note: This option only applies to the --cpu stressor option and not to all of the cpu class
of stressors.
--cpu-load-slice S
note: this option is only useful when --cpu-load is less than 100%. The CPU load is broken
into multiple busy and idle cycles. Use this option to specify the duration of a busy time
slice. A negative value for S specifies the number of iterations to run before idling the
CPU (e.g. -30 invokes 30 iterations of a CPU stress loop). A zero value selects a random
busy time between 0 and 0.5 seconds. A positive value for S specifies the number of
milliseconds to run before idling the CPU (e.g. 100 keeps the CPU busy for 0.1 seconds).
Specifying small values for S lends to small time slices and smoother scheduling. Setting
--cpu-load as a relatively low value and --cpu-load-slice to be large will cycle the CPU
between long idle and busy cycles and exercise different CPU frequencies. The thermal range
of the CPU is also cycled, so this is a good mechanism to exercise the scheduler, frequency
scaling and passive/active thermal cooling mechanisms.
Note: This option only applies to the --cpu stressor option and not to all of the cpu class
of stressors.
--cpu-method method
specify a cpu stress method. By default, all the stress methods are exercised sequentially,
however one can specify just one method to be used if required. Available cpu stress methods
are described as follows:
Method Description
all iterate over all the below cpu stress methods
ackermann Ackermann function: compute A(3, 7), where:
A(m, n) = n + 1 if m = 0;
A(m − 1, 1) if m > 0 and n = 0;
A(m − 1, A(m, n − 1)) if m > 0 and n > 0
For other recursive methods, refer to the hanoi cpu stress method.
apery calculate Apery's constant ζ(3); the sum of 1/(n ↑ 3) to a precision of
1.0x10↑14
bitops various bit operations from bithack, namely: reverse bits, parity check,
bit count, round to nearest power of 2
callfunc recursively call 8 argument C function to a depth of 1024 calls and unwind
cfloat 1000 iterations of a mix of floating point complex operations
cdouble 1000 iterations of a mix of double floating point complex operations
clongdouble 1000 iterations of a mix of long double floating point complex operations
collatz compute the 1348 steps in the collatz sequence starting from number
989345275647. Where f(n) = n / 2 (for even n) and f(n) = 3n + 1 (for odd
n).
correlate perform a 8192 × 512 correlation of random doubles
crc16 compute 1024 rounds of CCITT CRC16 on random data
decimal32 1000 iterations of a mix of 32 bit decimal floating point operations (GCC
only)
decimal64 1000 iterations of a mix of 64 bit decimal floating point operations (GCC
only)
decimal128 1000 iterations of a mix of 128 bit decimal floating point operations (GCC
only)
dither Floyd-Steinberg dithering of a 1024 × 768 random image from 8 bits down to
1 bit of depth
div8 50,000 8 bit unsigned integer divisions
div16 50,000 16 bit unsigned integer divisions
div32 50,000 32 bit unsigned integer divisions
div64 50,000 64 bit unsigned integer divisions
div128 50,000 128 bit unsigned integer divisions
double 1000 iterations of a mix of double precision floating point operations
euler compute e using n = (1 + (1 ÷ n)) ↑ n
explog iterate on n = exp(log(n) ÷ 1.00002)
factorial find factorials from 1..150 using Stirling's and Ramanujan's approximations
fibonacci compute Fibonacci sequence of 0, 1, 1, 2, 5, 8...
fft 4096 sample Fast Fourier Transform
fletcher16 1024 rounds of a naïve implementation of a 16 bit Fletcher's checksum
float 1000 iterations of a mix of floating point operations
float16 1000 iterations of a mix of 16 bit floating point operations
float32 1000 iterations of a mix of 32 bit floating point operations
float64 1000 iterations of a mix of 64 bit floating point operations
float80 1000 iterations of a mix of 80 bit floating point operations
float128 1000 iterations of a mix of 128 bit floating point operations
floatconversion perform 65536 iterations of floating point conversions between float,
double and long double floating point variables.
gamma calculate the Euler-Mascheroni constant γ using the limiting difference
between the harmonic series (1 + 1/2 + 1/3 + 1/4 + 1/5 ... + 1/n) and the
natural logarithm ln(n), for n = 80000.
gcd compute GCD of integers
gray calculate binary to gray code and gray code back to binary for integers
from 0 to 65535
hamming compute Hamming H(8,4) codes on 262144 lots of 4 bit data. This turns 4 bit
data into 8 bit Hamming code containing 4 parity bits. For data bits
d1..d4, parity bits are computed as:
p1 = d2 + d3 + d4
p2 = d1 + d3 + d4
p3 = d1 + d2 + d4
p4 = d1 + d2 + d3
hanoi solve a 21 disc Towers of Hanoi stack using the recursive solution. For
other recursive methods, refer to the ackermann cpu stress method.
hyperbolic compute sinh(θ) × cosh(θ) + sinh(2θ) + cosh(3θ) for float, double and long
double hyperbolic sine and cosine functions where θ = 0 to 2π in 1500 steps
idct 8 × 8 IDCT (Inverse Discrete Cosine Transform).
int8 1000 iterations of a mix of 8 bit integer operations.
int16 1000 iterations of a mix of 16 bit integer operations.
int32 1000 iterations of a mix of 32 bit integer operations.
int64 1000 iterations of a mix of 64 bit integer operations.
int128 1000 iterations of a mix of 128 bit integer operations (GCC only).
int32float 1000 iterations of a mix of 32 bit integer and floating point operations.
int32double 1000 iterations of a mix of 32 bit integer and double precision floating
point operations.
int32longdouble 1000 iterations of a mix of 32 bit integer and long double precision
floating point operations.
int64float 1000 iterations of a mix of 64 bit integer and floating point operations.
int64double 1000 iterations of a mix of 64 bit integer and double precision floating
point operations.
int64longdouble 1000 iterations of a mix of 64 bit integer and long double precision
floating point operations.
int128float 1000 iterations of a mix of 128 bit integer and floating point operations
(GCC only).
int128double 1000 iterations of a mix of 128 bit integer and double precision floating
point operations (GCC only).
int128longdouble 1000 iterations of a mix of 128 bit integer and long double precision
floating point operations (GCC only).
int128decimal32 1000 iterations of a mix of 128 bit integer and 32 bit decimal floating
point operations (GCC only).
int128decimal64 1000 iterations of a mix of 128 bit integer and 64 bit decimal floating
point operations (GCC only).
int128decimal128 1000 iterations of a mix of 128 bit integer and 128 bit decimal floating
point operations (GCC only).
intconversion perform 65536 iterations of integer conversions between int16, int32 and
int64 variables.
ipv4checksum compute 1024 rounds of the 16 bit ones' complement IPv4 checksum.
jmp Simple unoptimised compare >, <, == and jmp branching.
lfsr32 16384 iterations of a 32 bit Galois linear feedback shift register using
the polynomial x↑32 + x↑31 + x↑29 + x + 1. This generates a ring of 2↑32 −
1 unique values (all 32 bit values except for 0).
ln2 compute ln(2) based on series:
1 − 1/2 + 1/3 − 1/4 + 1/5 − 1/6 ...
logmap 16384 iterations computing chaotic double precision values using the
logistic map Χn+1 = r × Χn × (1 − Χn) where r > ∼ 3.56994567
longdouble 1000 iterations of a mix of long double precision floating point
operations.
loop simple empty loop.
matrixprod matrix product of two 128 × 128 matrices of double floats. Testing on 64
bit x86 hardware shows that this is provides a good mix of memory, cache
and floating point operations and is probably the best CPU method to use to
make a CPU run hot.
nsqrt compute sqrt() of long doubles using Newton-Raphson.
omega compute the omega constant defined by Ωe↑Ω = 1 using efficient iteration of
Ωn+1 = (1 + Ωn) / (1 + e↑Ωn).
parity compute parity using various methods from the Standford Bit Twiddling
Hacks. Methods employed are: the naïve way, the naïve way with the Brian
Kernigan bit counting optimisation, the multiply way, the parallel way, the
lookup table ways (2 variations) and using the __builtin_parity function.
phi compute the Golden Ratio ϕ using series.
pi compute π using the Srinivasa Ramanujan fast convergence algorithm.
prime find the first 10000 prime numbers using a slightly optimised brute force
naïve trial division search.
psi compute ψ (the reciprocal Fibonacci constant) using the sum of the
reciprocals of the Fibonacci numbers.
queens compute all the solutions of the classic 8 queens problem for board sizes
1..11.
rand 16384 iterations of rand(), where rand is the MWC pseudo random number
generator. The MWC random function concatenates two 16 bit
multiply-with-carry generators:
x(n) = 36969 × x(n − 1) + carry,
y(n) = 18000 × y(n − 1) + carry mod 2 ↑ 16
and has period of around 2 ↑ 60.
rand48 16384 iterations of drand48(3) and lrand48(3).
rgb convert RGB to YUV and back to RGB (CCIR 601).
sieve find the first 10000 prime numbers using the sieve of Eratosthenes.
stats calculate minimum, maximum, arithmetic mean, geometric mean, harmoninc mean
and standard deviation on 250 randomly generated positive double precision
values.
sqrt compute sqrt(rand()), where rand is the MWC pseudo random number generator.
trig compute sin(θ) × cos(θ) + sin(2θ) + cos(3θ) for float, double and long
double sine and cosine functions where θ = 0 to 2π in 1500 steps.
union perform integer arithmetic on a mix of bit fields in a C union. This
exercises how well the compiler and CPU can perform integer bit field loads
and stores.
zeta compute the Riemann Zeta function ζ(s) for s = 2.0..10.0
Note that some of these methods try to exercise the CPU with computations found in some real
world use cases. However, the code has not been optimised on a per-architecture basis, so may
be a sub-optimal compared to hand-optimised code used in some applications. They do try to
represent the typical instruction mixes found in these use cases.
--cpu-old-metrics
as of version V0.14.02 the cpu stressor now normalizes each of the cpu stressor method bogo-
op counters to try and ensure a similar bogo-op rate for all the methods to avoid the shorter
running (and faster) methods from skewing the bogo-op rates when using the default "all"
method. This is based on a reference Intel i5-8350U processor and hence the bogo-ops
normalizing factors will be skew somewhat on different CPUs, but so significantly as the
original bogo-op counter rates. To disable the normalization and fall back to the original
metrics, use this option.
--cpu-ops N
stop cpu stress workers after N bogo operations.
CPU onlining stressor
--cpu-online N
start N workers that put randomly selected CPUs offline and online. This Linux only stressor
requires root privilege to perform this action. By default the first CPU (CPU 0) is never
offlined as this has been found to be problematic on some systems and can result in a
shutdown.
--cpu-online-affinity
move the stressor worker to the CPU that will be next offlined.
--cpu-online-all
The default is to never offline the first CPU. This option will offline and online all the
CPUs including CPU 0. This may cause some systems to shutdown.
--cpu-online-ops N
stop after offline/online operations.
CPU affinity scheduler stressor
--cpu-sched N
start N workers that exercise the scheduler by moving processes to different CPUs. Each
worker starts 16 child processes and repeatedly moves the processes to different CPUs and
attempts changes their scheduler policy using SCHED_OTHER, SCHED_BATCH, SCHED_IDLE,
SCHED_EXT, SCHED_DEADLINE, SCHED_RR and SCHED_FIFO policies. The choice of CPU placement is
based on 8 different mechanisms and is changed every second to mix process placements on all
the available CPUS. The child processes are run with randomizied nice settings to exercise
scheduler prioritization.
--cpu-sched-ops N
stop after N child process move attempts.
Crypt stressor
--crypt N
start N workers that encrypt a 16 character random password using crypt(3). The password is
encrypted using MD5, NT, SHA-1, SHA-256, SHA-512, scrypt, SunMD5 and yescrypt encryption
methods.
--crypt-method method
select the encryption method, may be one of: all, MD5, NT, SHA-1, SHA-256, SHA-512, scrypt,
SunMD5 or yescrypt. The `all' method selects all the methods and is the default.
--crypt-ops N
stop after N bogo encryption operations.
Cyclic stressor
--cyclic N
start N workers that exercise the real time FIFO or Round Robin schedulers with cyclic
nanosecond sleeps. Normally one would just use 1 worker instance with this stressor to get
reliable statistics. By default this stressor measures the first 10 thousand latencies and
calculates the mean, mode, minimum, maximum latencies along with various latency percentiles
for the just the first cyclic stressor instance. One has to run this stressor with
CAP_SYS_NICE capability to enable the real time scheduling policies. The FIFO scheduling
policy is the default.
--cyclic-dist N
calculate and print a latency distribution with the interval of N nanoseconds. This is
helpful to see where the latencies are clustering.
--cyclic-method [ clock_ns | itimer | poll | posix_ns | pselect | usleep ]
specify the cyclic method to be used, the default is clock_ns. The available cyclic methods
are as follows:
Method Description
clock_ns sleep for the specified time using the clock_nanosleep(2) high resolution nanosleep
and the CLOCK_REALTIME real time clock.
itimer wakeup a paused process with a CLOCK_REALTIME itimer signal.
poll delay for the specified time using a poll delay loop that checks for time changes
using clock_gettime(2) on the CLOCK_REALTIME clock.
posix_ns sleep for the specified time using the POSIX nanosleep(2) high resolution
nanosleep.
pselect sleep for the specified time using pselect(2) with null file descriptors.
usleep sleep to the nearest microsecond using usleep(2).
--cyclic-ops N
stop after N sleeps.
--cyclic-policy [ deadline | fifo | rr ]
specify the desired real time scheduling policy, deadline, fifo (first-in, first-out) or rr
(round-robin).
--cyclic-prio P
specify the scheduling priority P. Range from 1 (lowest) to 100 (highest).
--cyclic-samples N
measure N samples. Range from 1 to 100000000 samples.
--cyclic-sleep N
sleep for N nanoseconds per test cycle using clock_nanosleep(2) with the CLOCK_REALTIME
timer. Range from 1 to 1000000000 nanoseconds.
Daemon stressor
--daemon N
start N workers that each create a daemon that dies immediately after creating another daemon
and so on. This effectively works through the process table with short lived processes that
do not have a parent and are waited for by init. This puts pressure on init to do rapid
child reaping. The daemon processes perform the usual mix of calls to turn into typical UNIX
daemons, so this artificially mimics very heavy daemon system stress.
--daemon-ops N
stop daemon workers after N daemons have been created.
--daemon-wait
wait for daemon child processes rather than let init handle the waiting. Enabling this option
will reduce the daemon fork rate because of the synchronous wait delays.
Datagram congestion control protocol (DCCP) stressor
--dccp N
start N workers that send and receive data using the Datagram Congestion Control Protocol
(DCCP) (RFC4340). This involves a pair of client/server processes performing rapid connect,
send and receives and disconnects on the local host.
--dccp-domain D
specify the domain to use, the default is ipv4. Currently ipv4 and ipv6 are supported.
--dccp-if NAME
use network interface NAME. If the interface NAME does not exist, is not up or does not
support the domain then the loopback (lo) interface is used as the default.
--dccp-msgs N
send N messages per connect, send/receive, disconnect iteration. The default is 10000
messages. If N is too small then the rate is throttled back by the overhead of dccp socket
connect and disconnects.
--dccp-port P
start DCCP at port P. For N dccp worker processes, ports P to P − 1 are used.
--dccp-ops N
stop dccp stress workers after N bogo operations.
--dccp-opts [ send | sendmsg | sendmmsg ]
by default, messages are sent using send(2). This option allows one to specify the sending
method using send(2), sendmsg(2) or sendmmsg(2). Note that sendmmsg is only available for
Linux systems that support this system call.
Mutex using Dekker algorithm stressor
--dekker N
start N workers that exercises mutex exclusion between two processes using shared memory with
the Dekker Algorithm. Where possible this uses memory fencing and falls back to using GCC
__sync_synchronize if they are not available. The stressors contain simple mutex and memory
coherency sanity checks.
--dekker-ops N
stop dekker workers after N mutex operations.
Dentry stressor
-D N, --dentry N
start N workers that create and remove directory entries. This should create file system
meta data activity. The directory entry names are suffixed by a gray-code encoded number to
try to mix up the hashing of the namespace.
--dentry-ops N
stop denty thrash workers after N bogo dentry operations.
--dentry-order [ forward | reverse | stride | random ]
specify unlink order of dentries, can be one of forward, reverse, stride or random. By
default, dentries are unlinked in random order. The forward order will unlink them from
first to last, reverse order will unlink them from last to first, stride order will unlink
them by stepping around order in a quasi-random pattern and random order will randomly select
one of forward, reverse or stride orders.
--dentries N
create N dentries per dentry thrashing loop, default is 2048.
/dev stressor
--dev N
start N workers that exercise the /dev devices. Each worker runs 5 concurrent threads that
perform open(2), fstat(2), lseek(2), poll(2), fcntl(2), mmap(2), munmap(2), fsync(2) and
close(2) on each device. Note that watchdog devices are not exercised.
--dev-file filename
specify the device file to exercise, for example, /dev/null. By default the stressor will
work through all the device files it can fine, however, this option allows a single device
file to be exercised.
--dev-ops N
stop dev workers after N bogo device exercising operations.
/dev/shm stressor
--dev-shm N
start N workers that fallocate large files in /dev/shm and then mmap these into memory and
touch all the pages. This exercises pages being moved to/from the buffer cache. Linux only.
--dev-shm-ops N
stop after N bogo allocation and mmap /dev/shm operations.
Decimal floating point operations stressor
--dfp N
start N workers that exercise addition, multiplication and division operations on a range of
decimal floating point types. For each type, 8 floating point values are operated upon 65536
times in a loop per bogo op.
--dfp-method method
select the decimal floating point method to use, available methods are:
Method Description
all iterate over all the following floating point methods:
df32add 32 bit decimal floating point add (_Decimal32)
df64add 64 bit decimal floating point add (_Decimal64)
df128add 128 bit decimal floating point add (_Decimal128)
df32mul 32 bit decimal floating point multiply (_Decimal32)
df64mul 64 bit decimal floating point multiply (_Decimal64)
df128mul 128 bit decimal floating point multiply (_Decimal128)
df32div 32 bit decimal floating point divide (_Decimal32)
df64div 64 bit decimal floating point divide (_Decimal64)
df128div 128 bit decimal floating point divide (_Decimal128)
Note that some of these decimal floating point methods may not be available on some systems.
--dfp-ops N
stop after N decimal floating point bogo ops.
Directories stressor
--dir N
start N workers that create, rename and remove directories using mkdir, rename and rmdir.
--dir-dirs N
exercise dir on N directories. The default is 8192 directories, this allows 64 to 65536
directories to be used instead.
--dir-ops N
stop directory thrash workers after N bogo directory operations.
Deep directories stressor
--dirdeep N
start N workers that create a depth-first tree of directories to a maximum depth as limited
by PATH_MAX or ENAMETOOLONG (which ever occurs first). By default, each level of the tree
contains one directory, but this can be increased to a maximum of 10 sub-trees using the
--dirdeep-dir option. To stress inode creation, a symlink and a hardlink to a file at the
root of the tree is created in each level.
--dirdeep-bytes N
allocated file size, the default is 0. One can specify the size as % of free space on the
file system or in units of Bytes, KBytes, MBytes and GBytes using the suffix b, k, m or g.
Used in conjunction with the --dirdeep-files option.
--dirdeep-dirs N
create N directories at each tree level. The default is just 1 but can be increased to a
maximum of 36 per level.
--dirdeep-files N
create N files at each tree level. The default is 0 with the file size specified by the
--dirdeep-bytes option.
--dirdeep-inodes N
consume up to N inodes per dirdeep stressor while creating directories and links. The value N
can be the number of inodes or a percentage of the total available free inodes on the
filesystem being used.
--dirdeep-ops N
stop directory depth workers after N bogo directory operations.
Maximum files creation in a directory stressor
--dirmany N
start N stressors that create as many files in a directory as possible and then remove them.
The file creation phase stops when an error occurs (for example, out of inodes, too many
files, quota reached, etc.) and then the files are removed. This cycles until the run time is
reached or the file creation count bogo-ops metric is reached. This is a much faster and
light weight directory exercising stressor compared to the dentry stressor.
--dirmany-bytes N
allocated file size, the default is 0. One can specify the size as % of free space on the
file system or in units of Bytes, KBytes, MBytes and GBytes using the suffix b, k, m or g.
--dirmany-ops N
stop dirmany stressors after N empty files have been created.
Dnotify stressor
--dnotify N
start N workers performing file system activities such as making/deleting files/directories,
renaming files, etc. to stress exercise the various dnotify events (Linux only).
--dnotify-ops N
stop inotify stress workers after N dnotify bogo operations.
Dup stressor
--dup N
start N workers that perform dup(2) and then close(2) operations on /dev/zero. The maximum
opens at one time is system defined, so the test will run up to this maximum, or 65536 open
file descriptors, which ever comes first.
--dup-ops N
stop the dup stress workers after N bogo open operations.
Dynamic libraries loading stressor
--dynlib N
start N workers that dynamically load and unload various shared libraries. This exercises
memory mapping and dynamic code loading and symbol lookups. See dlopen(3) for more details of
this mechanism.
--dynlib-ops N
stop workers after N bogo load/unload cycles.
Easy CPU opcode stressor
--easy-opcode N
start N workers that exercise 64 continuous pages of radomly selected simple CPU opcodes.
For example, for x86 processors this just exercises single byte opcodes such as nop and op-
codes that set/clear CPU flags to put pressure on the front-end decoder and instruction
cache. For RISC processors opcodes such as nop and simple register moves to/from the same
registers are used. Smart emulators may remove some of these opcodes (e.g move r0,r0) and
hence opcode rates may be artificially high.
--easy-opcode-ops N
stop after N executions of 64 continuous pages of opcodes.
Eigen C++ matrix library stressor
--eigen N
start N workers that exercise the Eigen C++ matrix library for 2D matrix addition,
multiplication, determinant, inverse and transpose operations on long double, double and
float matrices. This currently is only available for gcc/g++ builds.
--eigen-method method
select the floating point method to use, available methods are:
Method Description
all iterate over all the Eigen 2D matrix operations
add-longdouble addition of two matrices of long double floating point values
add-double addition of two matrices of double floating point values
add-float addition of two matrices of floating point values
determinant-longdouble determinant of matrix of long double floating point values
determinant-double determinant of matrix of double floating point values
determinant-float determinant of matrix of floating point values
inverse-longdouble inverse of matrix of long double floating point values
inverse-double inverse of matrix of double floating point values
inverse-float inverse of matrix of floating point values
multiply-longdouble multiplication of two matrices of long double floating point values
multiply-doublee multiplication of two matrices of double floating point values
multiply-float multiplication of two matrices of floating point values
transpose-longdouble transpose of matrix of long double floating point values
transpose-double transpose of matrix of double floating point values
transpose-float transpose of matrix of floating point values
--eigen-ops N
stop after N Eigen matrix computations
--eigen-size N
specify the 2D matrix size N × N. The default is a 32 × 32 matrix.
EFI variables stressor
--efivar N
start N workers that exercise the Linux /sys/firmware/efi/efivars and /sys/firmware/efi/vars
interfaces by reading the EFI variables. This is a Linux only stress test for platforms that
support the EFI vars interface and may require the CAP_SYS_ADMIN capability.
--efivar-ops N
stop the efivar stressors after N EFI variable read operations.
Non-functional system call (ENOSYS) stressor
--enosys N
start N workers that exercise non-functional system call numbers. This calls a wide range of
system call numbers to see if it can break a system where these are not wired up correctly.
It also keeps track of system calls that exist (ones that don't return ENOSYS) so that it can
focus on purely finding and exercising non-functional system calls. This stressor exercises
system calls from 0 to __NR_syscalls + 1024, random system calls within constrained in the
ranges of 0 to 2↑8, 2↑16, 2↑24, 2↑32, 2↑40, 2↑48, 2↑56 and 2↑64 bits, high system call
numbers and various other bit patterns to try to get wide coverage. To keep the environment
clean, each system call being tested runs in a child process with reduced capabilities.
--enosys-ops N
stop after N bogo enosys system call attempts
Environment variables stressor
--env N
start N workers that creates numerous large environment variables to try to trigger out of
memory conditions using setenv(3). If ENOMEM occurs then the environment is emptied and
another memory filling retry occurs. The process is restarted if it is killed by the Out Of
Memory (OOM) killer.
--env-ops N
stop after N bogo setenv/unsetenv attempts.
Epoll stressor
--epoll N
start N workers that perform various related socket stress activity using epoll_wait(2) to
monitor and handle new connections. This involves client/server processes performing rapid
connect, send/receives and disconnects on the local host. Using epoll allows a large number
of connections to be efficiently handled, however, this can lead to the connection table
filling up and blocking further socket connections, hence impacting on the epoll bogo op
stats. For ipv4 and ipv6 domains, multiple servers are spawned on multiple ports. The epoll
stressor is for Linux only.
--epoll-domain D
specify the domain to use, the default is unix (aka local). Currently ipv4, ipv6 and unix are
supported.
--epoll-ops N
stop epoll workers after N bogo operations.
--epoll-port P
start at socket port P. For N epoll worker processes, ports P to (P × 4) − 1 are used for
ipv4, ipv6 domains and ports P to P − 1 are used for the unix domain.
--epoll-sockets N
specify the maximum number of concurrently open sockets allowed in server. Setting a high
value impacts on memory usage and may trigger out of memory conditions.
Event file descriptor (eventfd) stressor
--eventfd N
start N parent and child worker processes that read and write 8 byte event messages between
them via the eventfd mechanism (Linux only).
--eventfd-nonblock
enable EFD_NONBLOCK to allow non-blocking on the event file descriptor. This will cause reads
and writes to return with EAGAIN rather the blocking and hence causing a high rate of polling
I/O.
--eventfd-ops N
stop eventfd workers after N bogo operations.
Exec processes stressor
--exec N
start N workers continually forking children that exec stress-ng and then exit almost
immediately. If a system has pthread support then 1 in 4 of the exec's will be from inside a
pthread to exercise exec'ing from inside a pthread context.
--exec-fork-method [ clone | fork | rfork | spawn | vfork ]
select the process creation method using clone(2), fork(2), BSD rfork(2), posix_spawn(3) or
vfork(2). Note that vfork will only exec programs using execve due to the constraints on the
shared stack between the parent and the child process.
--exec-max P
create P child processes that exec stress-ng and then wait for them to exit per iteration.
The default is 4096; higher values may create many temporary zombie processes that are
waiting to be reaped. One can potentially fill up the process table using high values for
--exec-max and --exec.
--exec-method [ all | execve | execveat ]
select the exec system call to use; all will perform a random choice between execve(2) and
execveat(2), execve will use execve(2) and execveat will use execveat(2) if it is available.
--exec-no-pthread
do not use pthread_create(3).
--exec-ops N
stop exec stress workers after N bogo operations.
Exiting pthread groups stressor
--exit-group N
start N workers that create 16 pthreads and terminate the pthreads and the controlling child
process using exit_group(2). (Linux only stressor).
--exit-group-ops N
stop after N iterations of pthread creation and deletion loops.
Exponential functions
--expmath N
start N workers that exercise various exponential functions with input values 0 to 1 in steps
of 0.001; the results are sanity checked to ensure no variation occurs after each round of
10000 computations.
--expmath-ops N
stop after N exponential bogo-operation loops.
--expmath-method method
specify a exponential function to exercise. Available exponential stress methods are
described as follows:
Method Description
all iterate over all the below exponential functions methods
cexp double complex natural exponential
cexpf float complex natural exponential
cexpl long double complex natural exponential
exp double natural exponential
expf float natural exponential
expl long double natural exponential
exp10 double base-10 exponential
exp10f float base-10 exponential
exp10l long double base-10 exponential
exp2 double base-2 exponential
exp2f float base-2 exponential
exp2l long double base-2 exponential
Factorization of large integers stressor
--factor N
start N workers that factorize large integers using the GNU Multiple Precision Arithmetic
Library. Randomized values to be factorized are computed so that an N digit value is
comprised of about 0.4 × N random factors, for N > 100. The default number of digits in the
value to be factorized is 10.
--factor-digits N
select the number of digits in the values to be factorized. Range 8 to 100000000 digits,
default is 10.
--factor-ops N
stop after N factorizations.
File space allocation (fallocate) stressor
-F N, --fallocate N
start N workers continually fallocating (preallocating file space) and ftruncating (file
truncating) temporary files. If the file is larger than the free space, fallocate will
produce an ENOSPC error which is ignored by this stressor.
--fallocate-bytes N
allocated file size, the default is 1 GB. One can specify the size as % of free space on the
file system or in units of Bytes, KBytes, MBytes and GBytes using the suffix b, k, m or g.
--fallocate-ops N
stop fallocate stress workers after N bogo fallocate operations.
Filesystem notification (fanotify) stressor
--fanotify N
start N workers performing file system activities such as creating, opening, writing, reading
and unlinking files to exercise the fanotify event monitoring interface (Linux only). Each
stressor runs a child process to generate file events and a parent process to read file
events using fanotify. Has to be run with CAP_SYS_ADMIN capability.
--fanotify-ops N
stop fanotify stress workers after N bogo fanotify events.
CPU branching instruction cache stressor
--far-branch N
start N workers that exercise calls to tens of thousands of functions that are relatively far
from the caller. All functions are 1 op instructions that return to the caller. The functions
are placed in pages that are memory mapped with a wide spread of fixed virtual addresses.
Function calls are pre-shuffled to create a randomized mix of addresses to call. This
stresses the instruction cache and any instruction TLBs.
--far-branch-flush
attempt to periodically flush instruction cache to produce instruction cache misses.
--far-branch-ops N
stop after N far branch bogo-ops. One full cycle of calling all the tens of thousands of
functions equates to one bogo-op.
--far-branch-pages N
specify the number of pages to allocate for far branch functions. The number for functions
per page depends on the processor architecture, for example, x86 will have 4096 x 1 byte
return instructions per 4 K page, where as SPARC64 will have only 512 x 8 byte return
instructions per 4 K page.
Page fault stressor
--fault N
start N workers that generates minor and major page faults.
--fault-ops N
stop the page fault workers after N bogo page fault operations.
Fcntl stressor
--fcntl N
start N workers that perform fcntl(2) calls with various commands. The exercised commands
(if available) are: F_CREATED_QUERY, F_DUPFD, F_DUPFD_CLOEXEC, F_GETFD, F_SETFD, F_GETFL,
F_SETFL, F_GETOWN, F_SETOWN, F_GETOWN_EX, F_SETOWN_EX, F_GETSIG, F_SETSIG, F_GETOWNER_UIDS,
F_GETLEASE, F_GETLK, F_SETLK, F_SETLKW, F_UNLCK, F_OFD_GETLK, F_OFD_SETLK, F_OFD_SETLKW,
F_GET_FILE_RW_HINT, F_SET_FILE_RW_HINT, F_GET_RW_HINT and F_SET_RW_HINT.
--fcntl-ops N
stop the fcntl workers after N bogo fcntl operations.
File descriptor abusing stressor
--fd-abuse N
start N workers that open file descriptors using various ways (file, dup, socket, pipe,
timerfd, pidfd, userfaultfd, etc.) and perform valid and invalid file operations on these
descriptors. This exercises a mix of successful and failed file operations on a wide range of
file descriptor types.
--fd-abuse-ops N
stop after N file abuse file operation bogo-ops.
File descriptor duplication and closing stressor
--fd-fork N
start N workers that open files using dup(2) on a file (/dev/zero by default) and then copies
these using multiple fork'd child processes and closes them with the fast clone_range(2) or
close(2) or by directly ending the processes using _exit(2). For every bogo-op, the stressor
attempts to dup(2) another 10000 file descriptors up to the maximum allowed, fork 8 child
processes that then close their copies of the file descriptors.
--fd-fork-fds N
specify maximum number of file descriptors to be opened. The default is 2 million, with a
range of 1000 to 16 million. The actual number used may be less depending on the system
defined limits of the number of open files per process.
--fd-fork-file [ null | random | stdin | stdout | zero ]
specify file to dup: null for /dev/null, random for /dev/random, stdin for standard input,
stdout for standard output, zero for /dev/zero. Default is /dev/zero.
--fd-fork-ops N
stop after N rounds of 10000 dups, forking/closing/exiting and waiting for the child
processes. Note that the bogo-ops metric rate will slow down over time as this stressor
increases the number of open files per bogo-loop and this increases the fork and close run
times.
File descriptor race stressor
--fd-race N
start N workers that attempt to force race conditions on opened file descriptors. Opened
file descriptors are passed from a server to a client over a socket. At periodic intervals
batches of the file descriptors are duplicated by creating multiple pthreads and then closed
en-masse with synchronized pthread termination. Also other concurrent pthreads exercise
various file based system calls on file descriptors that are in the process of being created.
By default a single file is used for the open calls, however /dev and /proc files can be
exercised using the appropriate fd-race options.
--fd-race-dev
exercise /dev files for race conditions.
--fd-race-race-ops N
stop after N file descriptors have been exercised.
--fd-race-proc
exercise /proc files for race conditions.
Fibonacci search stressor
--fibsearch N
start N workers that use a search a sorted array of 32 bit integers using a Fibonacci search.
A Fibonacci seaarch is similar to a bsearch except that it uses the Fibonacci series to
divide the search into unequal sized spaces. It avoids the costly division operator found in
bsearches and examines closer elements on each search step so there is a slight compute and
cache advantage over bsearch. By default, there are 65536 elements in the array. This is a
useful method to exercise random access of memory and processor cache.
--fibsearch-ops N
stop the fibsearch worker after N bogo fibsearch operations are completed.
--fibsearch-size N
specify the size (number of 32 bit integers) in the array to fibsearch. Size can be from 1 K
to 4 M.
File extent (fiemap) stressor
--fiemap N
start N workers that each create a file with many randomly changing extents and has 4 child
processes per worker that gather the extent information using the FS_IOC_FIEMAP ioctl(2).
--fiemap-bytes N
specify the size of the fiemap'd file in bytes. One can specify the size as % of free space
on the file system or in units of Bytes, KBytes, MBytes and GBytes using the suffix b, k, m
or g. Larger files will contain more extents, causing more stress when gathering extent
information.
--fiemap-ops N
stop after N fiemap bogo operations.
FIFO named pipe stressor
--fifo N
start N workers that exercise a named pipe by transmitting 64 bit integers.
--fifo-data-size N
set the byte size of the fifo write/reads, default is 8, range 8..4096.
--fifo-ops N
stop fifo workers after N bogo pipe write operations.
--fifo-readers N
for each worker, create N fifo reader workers that read the named pipe using simple blocking
reads. Default is 4, range 1..64.
File I/O control (ioctl) stressor
--file-ioctl N
start N workers that exercise various file specific ioctl(2) calls. This will attempt to use
the FIONBIO, FIOQSIZE, FIGETBSZ, FIOCLEX, FIONCLEX, FIONBIO, FIOASYNC, FIOQSIZE, FIFREEZE,
FITHAW, FICLONE, FICLONERANGE, FIONREAD, FIONWRITE and FS_IOC_RESVSP ioctls if these are
defined.
--file-ioctl-ops N
stop file-ioctl workers after N file ioctl bogo operations.
Filename stressor
--filename N
start N workers that exercise file creation using various length filenames containing a range
of allowed filename characters. This will try to see if it can exceed the file system
allowed filename length was well as test various filename lengths between 1 and the maximum
allowed by the file system.
--filename-ops N
stop filename workers after N bogo filename tests.
--filename-opts opt
use characters in the filename based on option `opt'. Valid options are:
Option Description
probe default option, probe the file system for valid allowed characters in a file name
and use these
posix use characters as specified by The Open Group Base Specifications Issue 7,
POSIX.1-2008, 3.278 Portable Filename Character Set
ext use characters allowed by the ext2, ext3, ext4 file systems, namely any 8 bit
character apart from NUL and /
File race stressor
--filerace N
start N workers that each run 8 concurrent processes that exercise a randomized set of file
related system calls on 64 random files. This attempts to trip any file system race
conditions while files are being operated upon. The randomized operations will cause some
invalid file operations to fail, these failures are silently ignored.
--filerace-ops N
stop after N randomized file bogo-operations, this is not a reliable metric of performance
because of the highly randomized actions occurring on randomly selected files.
Single cacheline coherency scalability stressor
--flipflop N
start N workers where each worker creates two groups of threads of the same size where each
group is affined to a set of CPUs. A continuous bitmap that has enough bits for each thread
pair is exercised, each thread pair tries to flip/flop their specific bit using cmpxchg
(compare/exchange); one thread tries to flip the bit from 0 to 1 and the other tries to flop
the bit from 1 to 0. This stressor makes threads compete on the same cacheline and measures
the total number of flip/flop operations and the distribution of successful flip/flops among
the threads (to see if thread pairs get starved in favour of others).
--flipflop-bits N
specifies number of bits in the bitmap (and hence number of flip/flop thread pairs).
--flipflop-taskset1 list
list of CPUs to affine the flip threads to. Refer to the --taskset option description for the
syntax of the list argument.
--flipflop-taskset2 list
list of CPUs to affine the flop threads to. Refer to the --taskset option description for the
syntax of the list argument.
--flipflop-ops N
stop after N bogo-ops, in this case a bogo-op is 100,000 flip-flop operations.
BSD File locking (flock) stressor
--flock N
start N workers locking on a single file.
--flock-ops N
stop flock stress workers after N bogo flock operations.
Cache flushing stressor
--flush-cache N
start N workers that flush the data and instruction cache (where possible). Some
architectures may not support cache flushing on either cache, in which case these become no-
ops.
--flush-cache-ops N
stop after N cache flush iterations.
Fused Multiply/Add floating point operations (fma) stressor
--fma N
start N workers that exercise single and double precision floating point multiplication and
add operations on arrays of 512 floating point values. More modern processors (Intel
Haswell, AMD Bulldozer and Piledriver) and modern C compilers these will be performed by
fused-multiply-add (fma3) opcodes. Operations used are:
a = (a × b) + c
a = (b × a) + c
a = (b × c) + a
a = (a × b) − c
a = (b × a) − c
a = (b × c) − a
--fma-libc
use libc fma math functions if they are available. These use either the libc FMA macros if
defined, the __builtin libc functions or the fma libc functions. Generally these are slower
than directly multiply/add fused code generated by the compiler.
--fma-ops N
stop after N bogo-loops of the 3 above operations on 512 single and double precision floating
point numbers.
Process forking stressor
-f N, --fork N
start N workers continually forking children that immediately exit.
--fork-max P
create P child processes and then wait for them to exit per iteration. The default is just 1;
higher values will create many temporary zombie processes that are waiting to be reaped. One
can potentially fill up the process table using high values for --fork-max and --fork.
--fork-ops N
stop fork stress workers after N bogo operations.
--fork-pageout
enable force paging-out of memory resident pages in fork stressor instances.
--fork-unmap
attempt to unmap unused non-memory resident shared library pages to try and reduced anonymous
vma copying. This is an ugly hack for benchmarking reduced vma copying and not guaranteed to
work. Linux only.
--fork-vm
enable detrimental performance virtual memory advice using madvise on all pages of the forked
process. Where possible this will try to set every page in the new process with using madvise
MADV_MERGEABLE, MADV_WILLNEED, MADV_HUGEPAGE and MADV_RANDOM flags. Linux only.
Heavy process forking stressor
--forkheavy N
start N workers that fork child processes from a parent that has thousands of allocated
system resources. The fork becomes a heavyweight operations as it has to duplicate the
resource references of the parent. Each stressor instance creates and reaps up to 4096 child
processes that are created and reaped in a first-in first-out manner.
--forkheavy-allocs N
attempt N resource allocation loops per stressor instance. Resources include pipes, file
descriptors, memory mappings, pthreads, timers, ptys, semaphores, message queues and
temporary files. These create heavyweight processes that are more time expensive to fork
from. Default is 16384.
--forkheavy-mlock
attempt to mlock future allocated pages into memory causing more memory pressure. If
mlock(MCL_FUTURE) is implemented then this will stop new brk pages from being swapped out.
--forkheavy-ops N
stop after N fork calls.
--forkheavy-procs N
attempt to fork N processes per stressor. The default is 4096 processes.
Floating point operations stressor
--fp N start N workers that exercise addition, multiplication and division operations on a range of
floating point types. For each type, 8 floating point values are operated upon 65536 times in
a loop per bogo op.
--fp-method method
select the floating point method to use, available methods are:
Method Description
all iterate over all the following floating point methods:
float128add 128 bit floating point add
ibm128add IBM 128 bit floating point add (powerpc)
float80add 80 bit floating point add
float64add 64 bit floating point add
float32add 32 bit binary32 floating point add
floatadd floating point add
bf16add bf16 floating point add
doubleadd double precision floating point add
ldoubleadd long double precision floating point add
float128mul 128 bit floating point multiply
ibm128mul IBM 128 bit floating point multiply (powerpc)
float80mul 80 bit floating point multiply
float64mul 64 bit floating point multiply
float32mul 32 bit binary32 floating point multiply
floatmul floating point multiply
bf16mul bf16 floating point multiply
doublemul double precision floating point multiply
ldoublemul long double precision floating point multiply
float128div 128 bit floating point divide
ibm128div IBM 128 bit floating point divide (powerpc)
float80div 80 bit floating point divide
float64div 64 bit floating point divide
float32div 32 bit binary32 floating point divide
floatdiv floating point divide
bf16div bf16 floating point divide
doublediv double precision floating point divide
ldoublediv long double precision floating point divide
Note that some of these floating point methods may not be available on some systems.
--fp-ops N
stop after N floating point bogo ops. Note that bogo-ops are counted for just standard float,
double and long double floating point types.
Floating point exception stressor
--fp-error N
start N workers that generate floating point exceptions. Computations are performed to force
and check for the FE_DIVBYZERO, FE_INEXACT, FE_INVALID, FE_OVERFLOW and FE_UNDERFLOW
exceptions. EDOM and ERANGE errors are also checked.
--fp-error-ops N
stop after N bogo floating point exceptions.
File punch and hole filling stressor
--fpunch N
start N workers that punch and fill holes in a 16 MB file using five concurrent processes per
stressor exercising on the same file. Where available, this uses fallocate(2)
FALLOC_FL_KEEP_SIZE, FALLOC_FL_PUNCH_HOLE, FALLOC_FL_ZERO_RANGE, FALLOC_FL_COLLAPSE_RANGE and
FALLOC_FL_INSERT_RANGE to make and fill holes across the file and breaks it into multiple
extents.
--fpunch-bytes N
set maximum size of each file for each fpunch worker process, the default is 16 MB. One can
specify the size as % of free space on the file system or in units of Bytes, KBytes, MBytes
and GBytes using the suffix b, k, m or g.
--fpunch-ops N
stop fpunch workers after N punch and fill bogo operations.
Fractal Stressor
--fractal N
start N workers that generate 2D fractals. By default a 1024 x 1024 point Mandelbrot set is
computed with a maximum of 256 iterations per point in the iterative compute loop. The
fractal is computed row by row with multiple rows shared amongst the N fractal stressor
instances. Double precision floating point values are used for the points in the complex set
of values. Naive computation is used with no special algorithmic short-cuts or interpolation.
--fractal-iterations N
specify the maximum number of iterations for the quadratic map computation of each point,
default is 256 iterations.
--fractal-method [ julia | mandelbrot ]
select the method of fractal generation, Julia set or Mandelbrot set, default is the
Mandelbrot set.
--fractal-ops N
stop after N fractals have been generated.
--fractal-sizex N
set the maximum width of the fractal, default is 1024 points.
--fractal-sizey N
set the maximum height of the fractal, default is 1024 points.
File size limit stressor
--fsize N
start N workers that exercise file size limits (via setrlimit RLIMIT_FSIZE) with file sizes
that are fixed, random and powers of 2. The files are truncated and allocated to trigger
SIGXFSZ signals.
--fsize-ops N
stop after N bogo file size test iterations.
File stats (fstat) stressor
--fstat N
start N workers fstat'ing files in a directory (default is /dev).
--fstat-dir directory
specify the directory to fstat to override the default of /dev. All the files in the
directory will be fstat'd repeatedly.
--fstat-ops N
stop fstat stress workers after N bogo fstat operations.
/dev/full stressor
--full N
start N workers that exercise /dev/full. This attempts to write to the device (which should
always get error ENOSPC), to read from the device (which should always return a buffer of
zeros) and to seek randomly on the device (which should always succeed). (Linux only).
--full-ops N
stop the stress full workers after N bogo I/O operations.
Function argument passing stressor
--funccall N
start N workers that call functions of 1 through to 9 arguments. By default all functions
with a range of argument types are called, however, this can be changed using the
--funccall-method option. This exercises stack function argument passing and re-ordering on
the stack and in registers.
--funccall-ops N
stop the funccall workers after N bogo function call operations. Each bogo operation is 1000
calls of functions of 1 through to 9 arguments of the chosen argument type.
--funccall-method method
specify the method of funccall argument type to be used. The default is all the types but can
be one of bool, uint8, uint16, uint32, uint64, uint128, float, double, longdouble, cfloat
(complex float), cdouble (complex double), clongdouble (complex long double), float16,
float32, float64, float80, float128, decimal32, decimal64 and decimal128. Note that some of
these types are only available with specific architectures and compiler versions.
Function return stressor
--funcret N
start N workers that pass and return by value various small to large data types.
--funcret-ops N
stop the funcret workers after N bogo function call operations.
--funcret-method method
specify the method of funcret argument type to be used. The default is uint64_t but can be
one of uint8 uint16 uint32 uint64 uint128 float double longdouble float80 float128 decimal32
decimal64 decimal128 uint8x32 uint8x128 uint64x128.
Fast mutex (futex) stressor
--futex N
start N workers that rapidly exercise the futex system call. Each worker has two processes, a
futex waiter and a futex waker. The waiter waits with a very small timeout to stress the
timeout and rapid polled futex waiting. This is a Linux specific stress option.
--futex-ops N
stop futex workers after N bogo successful futex wait operations.
Fetching data from kernel stressor
--get N
start N workers that call system calls that fetch data from the kernel, currently these are:
getpid, getppid, getcwd, getgid, getegid, getuid, getgroups, getpgrp, getpgid, getpriority,
getresgid, getresuid, getrlimit, prlimit, getrusage, getsid, gettid, getcpu, gettimeofday,
uname, adjtimex, sysfs. Some of these system calls are OS specific.
--get-ops N
stop get workers after N bogo get operations.
--get-slow-sync
attempt to synchronize system calls across the N get workers to try to force forms of locking
contention in the kernel on the more complex cases. Each system call is exercised
concurrently with the N workers for 0.1 seconds at a time, so it takes a 3-4 seconds to work
through all the system calls.
Virtual filesystem directories stressor (Linux)
--getdent N
start N workers that recursively read directories /proc, /dev/, /tmp, /sys and /run using
getdents and getdents64 (Linux only).
--getdent-ops N
stop getdent workers after N bogo getdent bogo operations.
Random data (getrandom) stressor
--getrandom N
start N workers that get 8192 random bytes from the /dev/urandom pool using the getrandom(2)
system call (Linux) or getentropy(2) (OpenBSD).
--getrandom-ops N
stop getrandom workers after N bogo get operations.
CPU pipeline and branch prediction stressor
--goto N
start N workers that perform 1024 forward branches (to next instruction) or backward branches
(to previous instruction) for each bogo operation loop. By default, every 1024 branches the
direction is randomly chosen to be forward or backward. This stressor exercises suboptimal
pipelined execution and branch prediction logic.
--goto-direction [ forward | backward | random ]
select the branching direction in the stressor loop, forward for forward only branching,
backward for a backward only branching, random for a random choice of forward or random
branching every 1024 branches.
--goto-ops N
stop goto workers after N bogo loops of 1024 branch instructions.
2D GPU stressor
--gpu N
start N worker that exercise the GPU. This specifies a 2-D texture image that allows the
elements of an image array to be read by shaders, and render primitives using an opengl
context.
--gpu-devnode DEVNAME
specify the device node name of the GPU device, the default is /dev/dri/renderD128.
--gpu-frag N
specify shader core usage per pixel, this sets N loops in the fragment shader.
--gpu-ops N
stop gpu workers after N render loop operations.
--gpu-tex-size N
specify upload texture N × N, by default this value is 4096 × 4096.
--gpu-xsize X
use a framebuffer size of X pixels. The default is 256 pixels.
--gpu-ysize Y
use a framebuffer size of Y pixels. The default is 256 pixels.
--gpu-upload N
specify upload texture N times per frame, the default value is 1.
Handle stressor
--handle N
start N workers that exercise the name_to_handle_at(2) and open_by_handle_at(2) system calls.
(Linux only).
--handle-ops N
stop after N handle bogo operations.
String hashing stressor
--hash N
start N workers that exercise various hashing functions. Random strings from 1 to 128 bytes
are hashed and the hashing rate and chi squared is calculated from the number of hashes
performed over a period of time. The chi squared value is the goodness-of-fit measure, it is
the actual distribution of items in hash buckets versus the expected distribution of items.
Typically a chi squared value of 0.95..1.05 indicates a good hash distribution.
--hash-method method
specify the hashing method to use, by default all the hashing methods are cycled through.
Methods available are:
Method Description
all cycle through all the hashing methods
adler32 Mark Adler checksum, a modification of the Fletcher checksum
coffin xor and 5 bit rotate left hash
coffin32 xor and 5 bit rotate left hash with 32 bit fetch optimization
crc32c compute CRC32C (Castagnoli CRC32) integer hash
djb2a Dan Bernstein hash using the xor variant
fnv1a FNV-1a Fowler-Noll-Vo hash using the xor then multiply variant
jenkin Jenkin's integer hash
kandr Kernighan and Richie's multiply by 31 and add hash from "The C Programming
Language", 2nd Edition
knuth Donald E. Knuth's hash from "The Art Of Computer Programming", Volume 3, chapter
6.4
loselose Kernighan and Richie's simple hash from "The C Programming Language", 1st Edition
mid5 xor shift hash of the middle 5 characters of the string. Designed by Colin Ian
King
muladd32 simple multiply and add hash using 32 bit math and xor folding of overflow
muladd64 simple multiply and add hash using 64 bit math and xor folding of overflow
mulxror32 32 bit multiply, xor and rotate right. Mangles 32 bits where possible. Designed
by Colin Ian King
mulxror64 64 bit multiply, xor and rotate right. 64 Bit version of mulxror32
murmur3_32 Austin Appleby's Murmur3 hash, 32 bit variant
nhash exim's nhash.
pjw a non-cryptographic hash function created by Peter J. Weinberger of AT&T Bell
Labs, used in UNIX ELF object files
sdbm sdbm hash as used in the SDBM database and GNU awk
sedgwick simple hash from Robert Sedgwick's C programming book
sobel Justin Sobel's bitwise shift hash
x17 multiply by 17 and add. The multiplication can be optimized down to a fast right
shift by 4 and add on some architectures
xor simple rotate shift and xor of values
xorror32 32 bit exclusive-or with right rotate hash, a fast string hash, designed by Colin
Ian King
xorror64 64 bit version of xorror32
xxhash the "Extremely fast" hash in non-streaming mode
--hash-ops N
stop after N hashing rounds
File-system stressor
-d N, --hdd N
start N workers continually writing, reading and removing temporary files. The default mode
is to stress test sequential writes and reads. With the --aggressive option enabled without
any --hdd-opts options the hdd stressor will work through all the --hdd-opt options one by
one to cover a range of I/O options.
--hdd-bytes N
write N bytes for each hdd process, the default is 1 GB. One can specify the size as % of
free space on the file system or in units of Bytes, KBytes, MBytes and GBytes using the
suffix b, k, m or g.
--hdd-opts list
specify various stress test options as a comma separated list. Options are as follows:
Option Description
direct try to minimize cache effects of the I/O. File I/O writes are performed
directly from user space buffers and synchronous transfer is also attempted.
To guarantee synchronous I/O, also use the sync option.
dsync ensure output has been transferred to underlying hardware and file metadata
has been updated (using the O_DSYNC open flag). This is equivalent to each
write(2) being followed by a call to fdatasync(2). See also the fdatasync
option.
fadv-dontneed advise kernel to expect the data will not be accessed in the near future.
fadv-noreuse advise kernel to expect the data to be accessed only once.
fadv-normal advise kernel there are no explicit access pattern for the data. This is the
default advice assumption.
fadv-rnd advise kernel to expect random access patterns for the data.
fadv-seq advise kernel to expect sequential access patterns for the data.
fadv-willneed advise kernel to expect the data to be accessed in the near future.
fsync flush all modified in-core data after each write to the output device using
an explicit fsync(2) call.
fdatasync similar to fsync, but do not flush the modified metadata unless metadata is
required for later data reads to be handled correctly. This uses an explicit
fdatasync(2) call.
iovec use readv/writev multiple buffer I/Os rather than read/write. Instead of 1
read/write operation, the buffer is broken into an iovec of 16 buffers.
noatime do not update the file last access timestamp, this can reduce metadata
writes.
sync ensure output has been transferred to underlying hardware (using the O_SYNC
open flag). This is equivalent to a each write(2) being followed by a call to
fsync(2). See also the fsync option.
rd-rnd read data randomly.
rd-seq read data sequentially.
syncfs write all buffered modifications of file metadata and data on the filesystem
that contains the hdd worker files.
utimes force update of file timestamp which may increase metadata writes.
wr-rnd write data randomly. The wr-seq option cannot be used at the same time.
wr-seq write data sequentially. This is the default if no write modes are specified.
Note that some of these options are mutually exclusive, for example, there can be only one method of
writing or reading. Also, fadvise flags may be mutually exclusive, for example fadv-willneed cannot
be used with fadv-dontneed.
--hdd-ops N
stop hdd stress workers after N bogo operations.
--hdd-write-size N
specify size of each write in bytes. Size can be from 1 byte to 4 MB.
BSD heapsort stressor
--heapsort N
start N workers that sort 32 bit integers using the BSD heapsort.
--heapsort-method [ heapsort-libc | heapsort-nonlibc ]
select either the libc implementation of heapsort or an optimized implementation of heapsort.
The default is the libc implementation if it is available.
--heapsort-ops N
stop heapsort stress workers after N bogo heapsorts.
--heapsort-size N
specify number of 32 bit integers to sort, default is 262144 (256 × 1024).
High resolution timer stressor
--hrtimers N
start N workers that exercise high resolution times at a high frequency. Each stressor starts
32 processes that run with random timer intervals of 0..499999 nanoseconds. Running this
stressor with appropriate privilege will run these with the SCHED_RR policy.
--hrtimers-adjust
enable automatic timer rate adjustment to try to maximize the hrtimer frequency. The signal
rate is measured every 0.1 seconds and the hrtimer delay is adjusted to try and set the
optimal hrtimer delay to generate the highest hrtimer rates.
--hrtimers-ops N
stop hrtimers stressors after N timer event bogo operations
Hashtable searching (hsearch) stressor
--hsearch N
start N workers that search a 80% full hash table using hsearch(3). By default, there are
8192 elements inserted into the hash table. This is a useful method to exercise access of
memory and processor cache.
--hsearch-method [ hsearch-libc | hsearch-nonlibc ]
select either the libc implementation of hsearch or a slightly optimized non-libc
implementation of hsearch. The default is the libc implementation if it exists, otherwise the
non-libc version.
--hsearch-ops N
stop the hsearch workers after N bogo hsearch operations are completed.
--hsearch-size N
specify the number of hash entries to be inserted into the hash table. Size can be from 1 K
to 4 M.
Hyperbolic functions stressor
--hyperbolic N
start N workers that exercise sinh, cosh, and tanh hyperbolic functions using float, double
and long double floating point variants. Each function is exercised 10,000 times per bogo-
operation.
--hyperbolic-method function
specify a hyperbolic stress function. By default, all the functions are exercised
sequentially, however one can specify just one function to be used if required. Available
options are as follows:
Method Description
all iterate through all of the following hyperbolic functions
cosh hyperbolic cosine (double precision)
coshf hyperbolic cosine (float precision)
coshl hyperbolic cosine (long double precision)
sinh hyperbolic sine (double precision)
sinhf hyperbolic sine (float precision)
sinhl hyperbolic sine (long double precision)
tanh hyperbolic tangent (double precision)
tanhf hyperbolic tangent (float precision)
tanhl hyperbolic tangent (long double precision)
--hyperbolic-ops N
stop after N bogo-operations.
CPU instruction cache load stressor
--icache N
start N workers that stress the instruction cache by forcing instruction cache reloads.
--icache-ops N
stop the icache workers after N bogo icache operations are completed.
ICMP flooding stressor
--icmp-flood N
start N workers that flood localhost with randomly sized ICMP ping packets. This stressor
requires the CAP_NET_RAW capbility.
--icmp-flood-max-size
use a maximum packet size of 65535 bytes instead of the default of 1000 bytes.
--icmp-flood-ops N
stop icmp flood workers after N ICMP ping packets have been sent.
Idle page scan stressor (Linux)
--idle-page N
start N workers that walks through every page exercising the Linux
/sys/kernel/mm/page_idle/bitmap interface. Requires CAP_SYS_RESOURCE capability.
--idle-page-ops N
stop after N bogo idle page operations.
Inode ioctl flags stressor
--inode-flags N
start N workers that exercise inode flags using the FS_IOC_GETFLAGS and FS_IOC_SETFLAGS
ioctl(2). This attempts to apply all the available inode flags onto a directory and file even
if the underlying file system may not support these flags (errors are just ignored). Each
worker runs 4 threads that exercise the flags on the same directory and file to try to force
races. This is a Linux only stressor, see ioctl_iflags(2) for more details.
--inode-flags-ops N
stop the inode-flags workers after N ioctl flag setting attempts.
Inotify stressor
--inotify N
start N workers performing file system activities such as making/deleting files/directories,
moving files, etc. to stress exercise the various inotify events (Linux only).
--inotify-ops N
stop inotify stress workers after N inotify bogo operations.
Insertion sort stressor
--insertionsort N
start N workers that sort 32 bit integers using insertion sort.
--insertionsort-ops N
stop insertionsort stress workers after N bogo insertion sorts.
--insertionsort-size N
specify number of 32 bit integers to sort, default is 16384 (16 × 1024).
Integer Math Operations
--intmath N
start N workers that perform addition, subtraction, multiplication, division and modulo math
operations on 128, 64, 32, 16 and 8 bit signed integers.
--intmath-fast
when available use int_fast64_t, int_fast32_t, int_fast16_t and int_fast8_t types instead of
int64_t, int32_t, int16_t and int8_t types. Note that these may or may not be faster than
normal integer operations depending on the compiler.
--intmath-method method
select the integer math method to use, available methods are:
Method Description
all iterate over all the following integer methods:
add128 128 bit signed integer addition
add64 64 bit signed integer addition
add32 32 bit signed integer addition
add16 16 bit signed integer addition
add8 8 bit signed integer addition
sub128 128 bit signed integer subtraction
sub64 64 bit signed integer subtraction
sub32 32 bit signed integer subtraction
sub16 16 bit signed integer subtraction
sub8 8 bit signed integer subtraction
mul128 128 bit signed integer multiplication
mul64 64 bit signed integer multiplication
mul32 32 bit signed integer multiplication
mul16 16 bit signed integer multiplication
mul8 8 bit signed integer multiplication
div128 128 bit signed integer division
div64 64 bit signed integer division
div32 32 bit signed integer division
div16 16 bit signed integer division
div8 8 bit signed integer division
mod128 128 bit signed integer modulo
mod64 64 bit signed integer modulo
mod32 32 bit signed integer modulo
mod16 16 bit signed integer modulo
mod8 8 bit signed integer modulo
for the --intmath-fast option, the following methods are available:
Method Description
all iterate over all the following integer methods:
addfast64 fast 64 bit signed integer addition
addfast32 fast 32 bit signed integer addition
addfast16 fast 16 bit signed integer addition
addfast8 fast 8 bit signed integer addition
subfast64 fast 64 bit signed integer subtraction
subfast32 fast 32 bit signed integer subtraction
subfast16 fast 16 bit signed integer subtraction
subfast8 fast 8 bit signed integer subtraction
mulfast64 fast 64 bit signed integer multiplication
mulfast32 fast 32 bit signed integer multiplication
mulfast16 fast 16 bit signed integer multiplication
mulfast8 fast 8 bit signed integer multiplication
divfast64 fast 64 bit signed integer division
divfast32 fast 32 bit signed integer division
divfast16 fast 16 bit signed integer division
divfast8 fast 8 bit signed integer division
modfast64 fast 64 bit signed integer modulo
modfast32 fast 32 bit signed integer modulo
modfast16 fast 16 bit signed integer modulo
modfast8 fast 8 bit signed integer modulo
--intmath-ops N
stop intmath workers after N bogo integer math operations.
Data synchronization (sync) stressor
-i N, --io N
start N workers continuously calling sync(2) to commit buffer cache to disk. This can be
used in conjunction with the --hdd stressor. This is a legacy stressor that is compatible
with the original stress tool.
--io-ops N
stop io stress workers after N bogo operations.
IO mixing stressor
--iomix N
start N workers that perform a mix of sequential, random and memory mapped read/write
operations as well as random copy file read/writes, forced sync'ing and (if run as root)
cache dropping. Multiple child processes are spawned to all share a single file and perform
different I/O operations on the same file.
--iomix-bytes N
write N bytes for each iomix worker process, the default is 1 GB. One can specify the size as
% of free space on the file system or in units of Bytes, KBytes, MBytes and GBytes using the
suffix b, k, m or g.
--iomix-ops N
stop iomix stress workers after N bogo iomix I/O operations.
Ioport stressor (x86 Linux)
--ioport N
start N workers than perform bursts of 16 reads and 16 writes of ioport 0x80 (x86 Linux
systems only). I/O performed on x86 platforms on port 0x80 will cause delays on the CPU
performing the I/O.
--ioport-ops N
stop the ioport stressors after N bogo I/O operations
--ioport-opts [ in | out | inout ]
specify the io operation to be performed. The default is both in and out. specify if port
reads in, port read writes out or reads and writes are
IO scheduling class and priority stressor
--ioprio N
start N workers that exercise the ioprio_get(2) and ioprio_set(2) system calls (Linux only).
--ioprio-ops N
stop after N io priority bogo operations.
Io-uring stressor
--io-uring N
start N workers that perform various io-uring file operations using the Linux io-uring
interface.
--io-uring-entries N
specify the number of io-uring ring entries.
--io-uring-ops
stop after N rounds of io-uring operations.
--io-uring-rand
randomize order of io-uring operations and file seek locations.
Ipsec multi-buffer cryptographic stressor
--ipsec-mb N
start N workers that perform cryptographic processing using the highly optimized Intel Multi-
Buffer Crypto for IPsec library. Depending on the features available, SSE, AVX, AVX and
AVX512 CPU features will be used on data encrypted by SHA, DES, CMAC, CTR, HMAC MD5, HMAC
SHA1 and HMAC SHA512 cryptographic routines. This is only available for x86-64 modern Intel
CPUs.
--ipsec-mb-feature [ sse | avx | avx2 | avx512 | noaesni ]
Just use the specified processor CPU feature. By default, all the available features for the
CPU are exercised.
--ipsec-mb-jobs N
Process N multi-block rounds of cryptographic processing per iteration. The default is 256.
--ipsec-mb-method [ all | cmac | ctr | des | hmac-md5 | hmac-sha1 | hmac-sha512 | sha ]
Select the ipsec-mb crypto/integrity method.
--ipsec-mb-ops N
stop after N rounds of processing of data using the cryptographic routines.
System interval timer stressor
--itimer N
start N workers that exercise the system interval timers. This sets up an ITIMER_PROF itimer
that generates a SIGPROF signal. The default frequency for the itimer is 1 MHz, however, the
Linux kernel will set this to be no more that the jiffy setting, hence high frequency SIGPROF
signals are not normally possible. A busy loop spins on getitimer(2) calls to consume CPU
and hence decrement the itimer based on amount of time spent in CPU and system time.
--itimer-freq F
run itimer at F Hz; range from 1 to 1000000 Hz. Normally the highest frequency is limited by
the number of jiffy ticks per second, so running above 1000 Hz is difficult to attain in
practice.
--itimer-ops N
stop itimer stress workers after N bogo itimer SIGPROF signals.
--itimer-rand
select an interval timer frequency based around the interval timer frequency +/- 12.5% random
jitter. This tries to force more variability in the timer interval to make the scheduling
less predictable.
Jpeg compression stressor
--jpeg N
start N workers that use jpeg compression on a machine generated plasma field image. The
default image is a plasma field, however different image types may be selected. The starting
raster line is changed on each compression iteration to cycle around the data.
--jpeg-height H
use a RGB sample image height of H pixels. The default is 512 pixels.
--jpeg-image [ brown | flat | gradient | noise | plasma | xstripes ]
select the source image type to be compressed. Available image types are:
Type Description
brown brown noise, red and green values vary by a 3 bit value, blue values vary by a 2
bit value.
flat a single random colour for the entire image.
gradient linear gradient of the red, green and blue components across the width and height
of the image.
noise random white noise for red, green, blue values.
plasma plasma field with smooth colour transitions and hard boundary edges.
xstripes a random colour for each horizontal line.
--jpeg-ops N
stop after N jpeg compression operations.
--jpeg-quality Q
use the compression quality Q. The range is 1..100 (1 lowest, 100 highest), with a default of
95
--jpeg-width H
use a RGB sample image width of H pixels. The default is 512 pixels.
Judy array stressor
--judy N
start N workers that insert, search and delete 32 bit integers in a Judy array using a
predictable yet sparse array index. By default, there are 131072 integers used in the Judy
array. This is a useful method to exercise random access of memory and processor cache.
--judy-ops N
stop the judy workers after N bogo judy operations are completed.
--judy-size N
specify the size (number of 32 bit integers) in the Judy array to exercise. Size can be from
1 K to 4 M 32 bit integers.
Kcmp stressor (Linux)
--kcmp N
start N workers that use kcmp(2) to compare parent and child processes to determine if they
share kernel resources. Supported only for Linux and requires CAP_SYS_PTRACE capability.
--kcmp-ops N
stop kcmp workers after N bogo kcmp operations.
Kernel key management stressor
--key N
start N workers that create and manipulate keys using add_key(2) and ketctl(2). As many keys
are created as the per user limit allows and then the following keyctl commands are exercised
on each key: KEYCTL_SET_TIMEOUT, KEYCTL_DESCRIBE, KEYCTL_UPDATE, KEYCTL_READ, KEYCTL_CLEAR
and KEYCTL_INVALIDATE.
--key-ops N
stop key workers after N bogo key operations.
Process signals stressor
--kill N
start N workers sending SIGUSR1 kill signals to a SIG_IGN signal handler in the stressor and
SIGUSR1 kill signal to a child stressor with a SIGUSR1 handler. Most of the process time will
end up in kernel space.
--kill-ops N
stop kill workers after N bogo kill operations.
Syslog stressor (Linux)
--klog N
start N workers exercising the kernel syslog(2) system call. This will attempt to read the
kernel log with various sized read buffers. Linux only.
--klog-ops N
stop klog workers after N syslog operations.
KVM stressor
--kvm N
start N workers that create, run and destroy a minimal virtual machine. The virtual machine
reads, increments and writes to port 0x80 in a spin loop and the stressor handles the I/O
transactions. Currently for x86 and Linux only.
--kvm-ops N
stop kvm stressors after N virtual machines have been created, run and destroyed.
CPU L1 cache stressor
--l1cache N
start N workers that exercise the CPU level 1 cache with reads and writes. A cache aligned
buffer that is twice the level 1 cache size is read and then written in level 1 cache set
sized steps over each level 1 cache set. This is designed to exercise cache block evictions.
The bogo-op count measures the number of million cache lines touched. Where possible, the
level 1 cache geometry is determined from the kernel, however, this is not possible on some
architectures or kernels, so one may need to specify these manually. One can specify 3 out of
the 4 cache geometric parameters, these are as follows:
--l1cache-line-size N
specify the level 1 cache line size (in bytes)
--l1cache-method [ forward | random | reverse ]
select the method of exercising a l1cache sized buffer. The default is a forward scan, random
picks random bytes to exercise, reverse scans in reverse.
--l1cache-mlock
attempt to mlock the l1cache size buffer into memory to prevent it from being swapped out.
--l1cache-ops N
specify the number of cache read/write bogo-op loops to run
--l1cache-sets N
specify the number of level 1 cache sets
--l1cache-size N
specify the level 1 cache size (in bytes)
--l1cache-ways N
specify the number of level 1 cache ways
Landlock stressor (Linux ≥ 5.13)
--landlock N
start N workers that exercise Linux 5.13 landlocking. A range of landlock_create_ruleset
flags are exercised with a read only file rule to see if a directory can be accessed and a
read-write file create can be blocked. Each ruleset attempt is exercised in a new child
context and this is the limiting factor on the speed of the stressor.
--landlock-ops N
stop the landlock stressors after N landlock ruleset bogo operations.
File lease stressor
--lease N
start N workers locking, unlocking and breaking leases via the fcntl(2) F_SETLEASE operation.
The parent processes continually lock and unlock a lease on a file while a user selectable
number of child processes open the file with a non-blocking open to generate SIGIO lease
breaking notifications to the parent. This stressor is only available if F_SETLEASE, F_WRLCK
and F_UNLCK support is provided by fcntl(2).
--lease-breakers N
start N lease breaker child processes per lease worker. Normally one child is plenty to
force many SIGIO lease breaking notification signals to the parent, however, this option
allows one to specify more child processes if required.
--lease-ops N
stop lease workers after N bogo operations.
LED stressor (Linux)
--led N
start N workers that exercise the /sys/class/leds interfaces to set LED brightness levels and
the various trigger settings. This needs to be run with root privilege to be able to write to
these settings successfully. Non-root privilege will ignore failed writes.
--led-ops N
stop after N interfaces are exercised.
Hardlink stressor
--link N
start N workers creating and removing hardlinks.
--link-ops N
stop link stress workers after N bogo operations.
--link-sync
sync dirty data and metadata to disk.
List data structures stressor
--list N
start N workers that exercise list data structures. The default is to add, find and remove
5,000 64 bit integers into circleq (doubly linked circle queue), list (doubly linked list),
slist (singly linked list), slistt (singly linked list using tail), stailq (singly linked
tail queue) and tailq (doubly linked tail queue) lists. The intention of this stressor is to
exercise memory and cache with the various list operations.
--list-method [ all | circleq | list | slist | stailq | tailq ]
specify the list to be used. By default, all the list methods are used (the `all' option).
--list-ops N
stop list stressors after N bogo ops. A bogo op covers the addition, finding and removing all
the items into the list(s).
--list-size N
specify the size of the list, where N is the number of 64 bit integers to be added into the
list.
Last level of cache stressor
--llc-affinity N
start N workers that exercise the last level of cache (LLC) by read/write activity across a
LLC sized buffer and then changing CPU affinity after each round of read/writes. This can
cause non-local memory stalls and LLC read/write misses.
--llc-affinity-clflush
where possible, flush cachelines after each cacheline write, x86 and ppc64 only.
--llc-affinity-mlock
attempt to mlock the LLC sized buffer into memory to prevent it from being swapped out.
--llc-affinity-numa
assign memory mapped pages to randomly selected NUMA nodes. This is disabled for systems that
do not support NUMA.
--llc-affinity-ops N
stop after N rounds of LLC read/writes.
Load average (loadavg) stressor
--loadavg N
start N workers that attempt to create thousands of pthreads that run at the lowest nice
priority to force very high load averages. Linux systems will also perform some I/O writes as
pending I/O is also factored into system load accounting.
--loadavg-max N
set the maximum number of pthreads to create to N. N may be reduced if there is as system
limit on the number of pthreads that can be created.
--loadavg-ops N
stop loadavg workers after N bogo scheduling yields by the pthreads have been reached.
Lock and increment memory stressor (x86 and ARM)
--lockbus N
start N workers that rapidly lock and increment 64 bytes of randomly chosen memory from a 16
MB mmap'd region (Intel x86 and ARM CPUs only). This will cause cacheline misses and
stalling of CPUs. Pages are spread randomly across NUMA nodes to exercise NUMA bus locking
for NUMA systems.
--lockbus-nosplit
disable split locks that lock across cache line boundaries.
--lockbus-ops N
stop lockbus workers after N bogo operations.
POSIX lock (F_SETLK/F_GETLK) stressor
--locka N
start N workers that randomly lock and unlock regions of a file using the POSIX advisory
locking mechanism (see fcntl(2), F_SETLK, F_GETLK). Each worker creates a 1024 KB file and
attempts to hold a maximum of 1024 concurrent locks with a child process that also tries to
hold 1024 concurrent locks. Old locks are unlocked in a first-in, first-out basis.
--locka-ops N
stop locka workers after N bogo locka operations.
POSIX lock (lockf) stressor
--lockf N
start N workers that randomly lock and unlock regions of a file using the POSIX lockf(3)
locking mechanism. Each worker creates a 64 KB file and attempts to hold a maximum of 1024
concurrent locks with a child process that also tries to hold 1024 concurrent locks. Old
locks are unlocked in a first-in, first-out basis.
--lockf-nonblock
instead of using blocking F_LOCK lockf(3) commands, use non-blocking F_TLOCK commands and re-
try if the lock failed. This creates extra system call overhead and CPU utilisation as the
number of lockf workers increases and should increase locking contention.
--lockf-ops N
stop lockf workers after N bogo lockf operations.
mixed file lock stressor (locka, lockf, lockofd)
--lockmix N
start N workers that randomly lock and unlock regions of a file using the BSD flock, locka
(advisory), POSIX lockf and Linux open file lock (lockofd) locking mechanisms. Each worker
creates a 1024 KB file and attempts to hold a maximum of 1024 concurrent locks with a child
process that also tries to hold 1024 concurrent locks. Old locks are unlocked in a first-in,
first-out basis.
--lockmix-ops N
stop lockmix workers after N bogo lockmix operations.
Linux open file lock (ofd F_OFD_SETLK/F_OFD_GETLK) stressor
--lockofd N
start N workers that randomly lock and unlock regions of a file using the Linux open file
description locks (see fcntl(2), F_OFD_SETLK, F_OFD_GETLK). Each worker creates a 1024 KB
file and attempts to hold a maximum of 1024 concurrent locks with a child process that also
tries to hold 1024 concurrent locks. Old locks are unlocked in a first-in, first-out basis.
--lockofd-ops N
stop lockofd workers after N bogo lockofd operations.
Logarithmic functions
--logmath N
start N workers that exercise various logarithmic functions with input values 1 to 10000.
Results are sanity checked to ensure no variation occurs after each round of 10000
computations.
--logmath-ops N
stop after N logarithmic bogo-operation loops.
--logmath-method method
specify a logarithmic function to exercise. Available logarithmic stress methods are
described as follows:
Method Description
all iterate over all the below logarithmic functions methods
clog double complex natural logarithm
clogf float complex natural logarithm
clogl long double complex natural logarithm
log double natural logarithm
logf float natural logarithm
logl long double natural logarithm
logb get exponent of a double
logbf get exponent of a float
logbl get exponent of a long double
log10 double base-10 logarithm
log10f float base-10 logarithm
log10l long double base-10 logarithm
log2 double base-2 logarithm
log2f float base-2 logarithm
log2l long double base-2 logarithm
Long jump (longjmp) stressor
--longjmp N
start N workers that exercise setjmp(3)/longjmp(3) by rapid looping on longjmp calls.
--longjmp-ops N
stop longjmp stress workers after N bogo longjmp operations (1 bogo op is 1000 longjmp
calls).
Loopback stressor (Linux)
--loop N
start N workers that exercise the loopback control device. This creates 2 MB loopback
devices, expands them to 4 MB, performs some loopback status information get and set
operations and then destroys them. Linux only and requires CAP_SYS_ADMIN capability.
--loop-ops N
stop after N bogo loopback creation/deletion operations.
Linear search stressor
--lsearch N
start N workers that linear search a unsorted array of 32 bit integers using lsearch(3). By
default, there are 8192 elements in the array. This is a useful method to exercise
sequential access of memory and processor cache.
--lsearch-method [ lsearch-libc | lsearch-nonlibc | lsearch-sentinel ]
select either the libc implementation of lsearch or a slightly optimized non-libc
implementation of lsearch or a lsearch that uses the search key as the end of array sentinel
to remove an index compare per loop. The default is the libc implementation if it exists,
otherwise the non-libc version.
--lsearch-ops N
stop the lsearch workers after N bogo lsearch operations are completed.
--lsearch-size N
specify the size (number of 32 bit integers) in the array to lsearch. Size can be from 1 K to
4 M.
Linux Security Modules system call stressor
--lsm N
start N workers that exercise the LSM system calls lsm_list_modules and lsm_get_self_attr,
(Linux only). Each bogo-op loop fetches a list of available security modules and fetching
LSM attributes as well as some invalid LSM system calls to exercise error handling.
--lsm-ops N
stop after N loops of fetching security modules lists and fetching LSM attributes.
Madvise stressor
--madvise N
start N workers that apply random madvise(2) advise settings on pages of a 4 MB file backed
shared memory mapping.
--madvise-hwpoison
enable MADV_HWPOISON page poisoning (if available, only when run as root). This will page
poison a few pages and will cause kernel error messages to be reported.
--madvise-ops N
stop madvise stressors after N bogo madvise operations.
Memory allocation stressor
--malloc N
start N workers continuously calling malloc(3), calloc(3), realloc(3), posix_memalign(3),
aligned_alloc(3), memalign(3) and free(3). By default, up to 65536 allocations can be active
at any point, but this can be altered with the --malloc-max option. Allocation, reallocation
and freeing are chosen at random; 50% of the time memory is allocation (via one of malloc,
calloc or realloc, posix_memalign, aligned_alloc, memalign) and 50% of the time allocations
are free'd. Allocation sizes are also random, with the maximum allocation size controlled by
the --malloc-bytes option, the default size being 64 K. The worker is re-started if it is
killed by the out of memory (OOM) killer.
--malloc-bytes N
maximum per allocation/reallocation size. Allocations are randomly selected from 1 to N
bytes. One can specify the size as % of total available memory or in units of Bytes, KBytes,
MBytes and GBytes using the suffix b, k, m or g. Large allocation sizes cause the memory
allocator to use mmap(2) rather than expanding the heap using brk(2).
--malloc-max N
maximum number of active allocations allowed. Allocations are chosen at random and placed in
an allocation slot. Because about 50%/50% split between allocation and freeing, typically
half of the allocation slots are in use at any one time.
--malloc-mlock
attempt to mlock the allocations into memory to prevent them from being swapped out.
--malloc-ops N
stop after N malloc bogo operations. One bogo operations relates to a successful malloc(3),
calloc(3), realloc(3), posix_memalign(3), aligned_alloc(3) or memalign(3) call.
--malloc-pthreads N
specify number of malloc stressing concurrent pthreads to run. The default is 0 (just one
main process, no pthreads). This option will do nothing if pthreads are not supported.
--malloc-thresh N
specify the threshold where malloc uses mmap(2) instead of sbrk(2) to allocate more memory.
This is only available on systems that provide the GNU C mallopt(3) tuning function.
--malloc-touch
touch every allocated page to force pages to be populated in memory. This will increase the
memory pressure and exercise the virtual memory harder. By default the malloc stressor will
madvise pages into memory or use mincore to check for non-resident memory pages and try to
force them into memory; this option aggressively forces pages to be memory resident.
--malloc-trim
periodically trim memory allocation by attempting to release free memory from the heap every
65536 allocation iterations. This can be a time consuming operation. It is only available
with libc malloc implementations that support malloc_trim(3).
--malloc-zerofree
zero allocated memory before free'ing. This can be useful in touching broken allocations and
triggering failures. Also useful for forcing extra cache/memory writes.
2D Matrix stressor
--matrix N
start N workers that perform various matrix operations on floating point values. Testing on
64 bit x86 hardware shows that this provides a good mix of memory, cache and floating point
operations and is an excellent way to make a CPU run hot.
By default, this will exercise all the matrix stress methods one by one on a 128 × 128
element matrix. One can specify a specific matrix stress method with the --matrix-method
option.
--matrix-method method
specify a matrix stress method. Available matrix stress methods are described as follows:
Method Description
all iterate over all the below matrix stress methods
add add two N × N matrices
copy copy one N × N matrix to another
div divide an N × N matrix by a scalar
frobenius Frobenius product of two N × N matrices
hadamard Hadamard product of two N × N matrices
identity create an N × N identity matrix
mean arithmetic mean of two N × N matrices
mult multiply an N × N matrix by a scalar
negate negate an N × N matrix
prod product of two N × N matrices
sub subtract one N × N matrix from another N × N matrix
square multiply an N × N matrix by itself
trans transpose an N × N matrix
zero zero an N × N matrix
--matrix-ops N
stop matrix stress workers after N bogo operations.
--matrix-size N
specify the N × N size of the matrices. Smaller values result in a floating point compute
throughput bound stressor, where as large values result in a cache and/or memory bandwidth
bound stressor.
--matrix-yx
perform matrix operations in order y by x rather than the default x by y. This is suboptimal
ordering compared to the default and will perform more data cache stalls.
3D Matrix stressor
--matrix-3d N
start N workers that perform various 3D matrix operations on floating point values. Testing
on 64 bit x86 hardware shows that this provides a good mix of memory, cache and floating
point operations and is an excellent way to make a CPU run hot.
By default, this will exercise all the 3D matrix stress methods one by one on a 128 × 128 ×
128 element matrix. One can specify a specific 3D matrix stress method with the
--matrix-3d-method option.
--matrix-3d-method method
specify a 3D matrix stress method. Available 3D matrix stress methods are described as
follows:
Method Description
all iterate over all the below matrix stress methods
add add two N × N × N matrices
copy copy one N × N × N matrix to another
div divide an N × N × N matrix by a scalar
frobenius Frobenius product of two N × N × N matrices
hadamard Hadamard product of two N × N × N matrices
identity create an N × N × N identity matrix
mean arithmetic mean of two N × N × N matrices
mult multiply an N × N × N matrix by a scalar
negate negate an N × N × N matrix
sub subtract one N × N × N matrix from another N × N × N matrix
trans transpose an N × N × N matrix
zero zero an N × N × N matrix
--matrix-3d-ops N
stop the 3D matrix stress workers after N bogo operations.
--matrix-3d-size N
specify the N × N × N size of the matrices. Smaller values result in a floating point
compute throughput bound stressor, where as large values result in a cache and/or memory
bandwidth bound stressor.
--matrix-3d-zyx
perform matrix operations in order z by y by x rather than the default x by y by z. This is
suboptimal ordering compared to the default and will perform more data cache stalls.
Memory contention stressor
--mcontend N
start N workers that produce memory contention read/write patterns. Each stressor runs with 5
threads that read and write to two different mappings of the same underlying physical page.
Various caching operations are also exercised to cause sub-optimal memory access patterns.
The threads also randomly change CPU affinity to exercise CPU and memory migration stress.
--mcontend-numa
assign memory mapped pages to randomly selected NUMA nodes. This is disabled for systems that
do not support NUMA.
--mcontend-ops N
stop mcontend stressors after N bogo read/write operations.
Memory barrier stressor (Linux)
--membarrier N
start N workers that exercise the membarrier system call (Linux only).
--membarrier-ops N
stop membarrier stress workers after N bogo membarrier operations.
Memory copy (memcpy) stressor
--memcpy N
start N workers that copies data to and from a buffer using memcpy(3) and then move the data
in the buffer with memmove(3) with 3 different alignments. This will exercise the data cache
and memory copying.
--memcpy-method [ all | libc | builtin | naive | naive_o0 .. naive_o3 ]
specify a memcpy copying method. Available memcpy methods are described as follows:
Method Description
all use libc, builtin and naïve methods
libc use libc memcpy and memmove functions, this is the default
builtin use the compiler built in optimized memcpy and memmove functions
naive use naïve byte by byte copying and memory moving build with default compiler
optimization flags
naive_o0 use unoptimized naïve byte by byte copying and memory moving
naive_o1 use unoptimized naïve byte by byte copying and memory moving with -O1 optimization
naive_o2 use optimized naïve byte by byte copying and memory moving build with -O2
optimization and where possible use CPU specific optimizations
naive_o3 use optimized naïve byte by byte copying and memory moving build with -O3
optimization and where possible use CPU specific optimizations
--memcpy-ops N
stop memcpy stress workers after N bogo memcpy operations.
Anonymous file (memfd) stressor
--memfd N
start N workers that create allocations of 1024 pages using memfd_create(2) and ftruncate(2)
for allocation and mmap(2) to map the allocation into the process address space. (Linux
only).
--memfd-bytes N
allocate N bytes per memfd stress worker, the default is 256 MB. One can specify the size in
as % of total available memory or in units of Bytes, KBytes, MBytes and GBytes using the
suffix b, k, m or g.
--memfd-fds N
create N memfd file descriptors, the default is 256. One can select 8 to 4096 memfd file
descriptions with this option.
--memfd-madvise
enable random madvise page advice on memfd memory mapped regions to add a little more VM
exercising.
--memfd-mlock
attempt to mlock mmap'd pages into memory causing more memory pressure by preventing pages
from swapped out.
--memfd-numa
assign memory mapped pages to randomly selected NUMA nodes. This is disabled for systems that
do not support NUMA.
--memfd-ops N
stop after N memfd-create(2) bogo operations.
--memfd-zap-pte
exercise zapping page-table-entries to try to reproduce a Linux kernel bug that was fixed by
commit 5abfd71d936a8aefd9f9ccd299dea7a164a5d455 "mm: don't skip swap entry even if
zap_details specified". This will slow the stressor down significantly and hence is an opt-in
memfd stressor option.
Memory hotplug stressor (Linux)
--memhotplug N
start N workers that offline and online memory hotplug regions. Linux only and requires
CAP_SYS_ADMIN capabilities.
--memhotplug-mmap
enable random 1 K to 1 MB memory mapping/unmappings before each offline event.
--memhotplug-ops N
stop memhotplug stressors after N memory offline and online bogo operations.
Memory read/write stressor
--memrate N
start N workers that exercise a buffer with 1024, 512, 256, 128, 64, 32, 16 and 8 bit reads
and writes. 1024, 512 and 256 reads and writes are available with compilers that support
integer vectors. x86-64 cpus that support uncached (non-temporal "nt") writes also exercise
128, 64 and 32 writes providing higher write rates than the normal cached writes. x86-64 also
exercises repeated string stores using 64, 32, 16 and 8 bit writes. CPUs that support
prefetching reads also exercise 64 prefetched "pf" reads. This memory stressor allows one to
also specify the maximum read and write rates. The stressors will run at maximum speed if no
read or write rates are specified.
--memrate-bytes N
specify the size of the memory buffer being exercised. The default size is 256 MB. One can
specify the size in units of Bytes, KBytes, MBytes and GBytes using the suffix b, k, m or g,
or cache sizes with L1, L2, L3 or LLC (lower level cache size).
--memrate-flush
flush cache between each memory exercising test to remove caching benefits in memory rate
metrics.
--memrate-method
specify a memrate stress method, some methods are available to specific architectures or
toolchains that support them. Available memrate stress methods are described as follows:
Method Description
all iterate over all the below memrate methods
read128pf read 128 bits per read using prefetching
read64pf read 64 bits per read using prefetching
read1024 read a vector of 1024 bits per read
read512 read a vector of 512 bits per read
read256 read a vector of 256 bits per read
read128 read 128 bits per read
read64 read 64 bits per read
read32 read 32 bits per read
read16 read 16 bits per read
read8 read 8 bits per read
write64stoq write 64 bits per write with x86 rep stoq
write32stow write 32 bits per write with x86 rep stow
write16stod write 16 bits per write with x86 rep stod
write8stob write 8 bits per write with x86 rep stob
write64ds write 64 bits per write with x86 movdiri
write128nt write 128 bits per write using non-temporal store
write64nt write 64 bits per write using non-temporal store
write32nt write 32 bits per write using non-temporal store
write1024 write a vector of 1024 bits per write
write512 write a vector of 512 bits per write
write256 write a vector of 256 bits per write
write128 write 128 bits per write
write64 write 64 bits per write
write32 write 32 bits per write
write16 write 16 bits per write
write8 write 8 bits per write
memset write using libc memset
--memrate-ops N
stop after N bogo memrate operations.
--memrate-rd-mbs N
specify the maximum allowed read rate in MB/sec. The actual read rate is dependent on
scheduling jitter and memory accesses from other running processes. Setting this to zero will
disable reads.
--memrate-wr-mbs N
specify the maximum allowed read rate in MB/sec. The actual write rate is dependent on
scheduling jitter and memory accesses from other running processes. Setting this to zero will
disable writes.
Memory thrash stressor
--memthrash N
start N workers that thrash and exercise a 16 MB buffer in various ways to try and trip
thermal overrun. Each stressor will start 1 or more threads. The number of threads is
chosen so that there will be at least 1 thread per CPU. Note that the optimal choice for N is
a value that divides into the number of CPUs.
--memthrash-method method
specify a memthrash stress method. Available memthrash stress methods are described as
follows:
Method Description
all iterate over all the below memthrash methods
chunk1 memset 1 byte chunks of random data into random locations
chunk8 memset 8 byte chunks of random data into random locations
chunk64 memset 64 byte chunks of random data into random locations
chunk256 memset 256 byte chunks of random data into random locations
chunkpage memset page size chunks of random data into random locations
copy128 copy 128 byte chunks from chunk N + 1 to chunk N with streaming reads and writes
with 128 bit memory accesses where possible.
flip flip (invert) all bits in random locations
flush flush cache line in random locations
lock lock randomly choosing locations (Intel x86 and ARM CPUs only)
matrix treat memory as a 2 × 2 matrix and swap random elements
memmove copy all the data in buffer to the next memory location
memset memset the memory with random data
memset64 memset the memory with a random 64 bit value in 64 byte chunks using non-
temporal stores if possible or normal stores as a fallback
memsetstosd memset the memory using x86 32 bit rep stosd instruction (x86 only)
mfence stores with write serialization
numa memory bind pages across numa nodes
prefetch prefetch data at random memory locations
random randomly run any of the memthrash methods except for `random' and `all'
reverse swap 8 bit values from start to end and work towards the middle
spinread spin loop read the same random location 2↑19 times
spinwrite spin loop write the same random location 2↑19 times
swap step through memory swapping bytes in steps of 65 and 129 byte strides
swap64 work through memory swapping adjacent 64 byte chunks
swapfwdrev swap 64 bit values from start to end and work towards the middle and then from
end to start and work towards the middle.
tlb work through memory in sub-optimial strides of prime multiples of the cache line
size with reads and then writes to cause Translation Lookaside Buffer (TLB)
misses.
--memthrash-ops N
stop after N memthrash bogo operations.
BSD mergesort stressor
--mergesort N
start N workers that sort 32 bit integers using the BSD mergesort.
--mergesort-method [ mergesort-libc | mergesort-nonlibc ]
select either the libc implementation of mergesort or an unoptimized implementation of
mergesort. The default is the libc implementation if it is available.
--mergesort-ops N
stop mergesort stress workers after N bogo mergesorts.
--mergesort-size N
specify number of 32 bit integers to sort, default is 262144 (256 × 1024).
File metadata mix
--metamix N
start N workers that generate a file metadata mix of operations. Each stressor runs 16
concurrent processes that each exercise a file's metadata with sequences of open, 256 lseeks
and writes, fdatasync, close, fsync and then stat, open, 256 lseeks, reads, occasional file
memory mapping, close, unlink and lstat.
--metamix-bytes N
set the size of metamix files, the default is 1 MB. One can specify the size as % of free
space on the file system or in units of Bytes, KBytes, MBytes and GBytes using the suffix b,
k, m or g.
--metamix-ops N
stop the metamix stressor after N bogo metafile operations.
Resident memory (mincore) stressor
--mincore N
start N workers that walk through all of memory 1 page at a time checking if the page mapped
and also is resident in memory using mincore(2). It also maps and unmaps a page to check if
the page is mapped or not using mincore(2).
--mincore-ops N
stop after N mincore bogo operations. One mincore bogo op is equivalent to a 300 mincore(2)
calls.
--mincore-random
instead of walking through pages sequentially, select pages at random. The chosen address is
iterated over by shifting it right one place and checked by mincore until the address is less
or equal to the page size.
Minimum sleep time in nanosleep stressor
--min-nanosleep N
start M workers that exercise nanosecond sleeps using powers of two nanosecond sleep delays.
Once all the instances have completed, the minimum, maximum and mean sleep times are reported
for the sleep delays across all the min-nanosleep stressor instances.
--min-nanosleep-ops N
stop after N rounds of measurements across all the sleeps are completed.
--min-nanosleep-max N
set the maximum nanosleep delay to use. If this is not a power of two then the previous power
of two nanosecond delay time is used, e.g. specifying 10000 will select 8192 nanoseconds.
--min-nanosleep-sched [ batch | deadline | fifo | idle | other | rr ]
select scheduling policy. Note that deadline, fifo and rr require root privilege.
Misaligned read/write stressor
--misaligned N
start N workers that perform misaligned read and writes. By default, this will exercise 128
bit misaligned read and writes in 8 × 16 bits, 4 × 32 bits, 2 × 64 bits and 1 × 128 bits at
the start of a page boundary, at the end of a page boundary and over a cache boundary.
Misaligned read and writes operate at 1 byte offset from the natural alignment of the data
type. On some architectures this can cause SIGBUS, SIGILL or SIGSEGV, these are handled and
the misaligned stressor method causing the error is disabled.
--misaligned-method method
Available misaligned stress methods are described as follows:
Method Description
all iterate over all the following misaligned methods
int16rd 8 × 16 bit integer reads
int16wr 8 × 16 bit integer writes
int16inc 8 × 16 bit integer increments
int16atomic 8 × 16 bit atomic integer increments
int32rd 4 × 32 bit integer reads
int32wr 4 × 32 bit integer writes
int32wtnt 4 × 32 bit non-temporal stores (x86 only)
int32inc 4 × 32 bit integer increments
int32atomic 4 × 32 bit atomic integer increments
int64rd 2 × 64 bit integer reads
int64wr 2 × 64 bit integer writes
int64wrds 4 × 64 bit direct stores (x86 only)
int64wtnt 4 × 64 bit non-temporal stores (x86 only)
int64inc 2 × 64 bit integer increments
int64atomic 2 × 64 bit atomic integer increments
int128rd 1 × 128 bit integer reads
int128wr 1 × 128 bit integer writes
int128inc 1 × 128 bit integer increments
int128atomic 1 × 128 bit atomic integer increments
Note that some of these options (128 bit integer and/or atomic operations) may not be available on
some systems.
--misaligned-ops N
stop after N misaligned bogo operation. A misaligned bogo op is equivalent to 65536 × 128 bit
reads or writes.
Mknod/unlink stressor
--mknod N
start N workers that create and remove fifos, empty files and named sockets using mknod and
unlink.
--mknod-ops N
stop directory thrash workers after N bogo mknod operations.
Memory mapped pages lock/unlock stressor
--mlock N
start N workers that lock and unlock memory mapped pages using mlock(2), munlock(2),
mlockall(2) and munlockall(2). This is achieved by the mapping of three contiguous pages and
then locking the second page, hence ensuring non-contiguous pages are locked . This is then
repeated until the maximum allowed mlocks or a maximum of 262144 mappings are made. Next,
all future mappings are mlocked and the worker attempts to map 262144 pages, then all pages
are munlocked and the pages are unmapped.
--mlock-ops N
stop after N mlock bogo operations.
Many child memory mapped page lock/unlock process stressor
--mlockmany N
start N workers that fork off a default of 1024 child processes in total; each child will
attempt to anonymously mmap and mlock the maximum allowed mlockable memory size. The stress
test attempts to avoid swapping by tracking low memory and swap allocations (but some
swapping may occur). Once either the maximum number of child process is reached or all
mlockable in-core memory is locked then child processes are killed and the stress test is
repeated.
--mlockmany-ops N
stop after N mlockmany (mmap and mlock) operations.
--mlockmany-procs N
set the number of child processes to create per stressor. The default is to start a maximum
of 1024 child processes in total across all the stressors. This option allows the setting of
N child processes per stressor.
Memory mapping (mmap/munmap) stressor
--mmap N
start N workers continuously calling mmap(2)/munmap(2). The initial mapping is a large chunk
(size specified by --mmap-bytes) followed by pseudo-random 4 K unmappings, then pseudo-random
4 K mappings, and then linear 4 K unmappings. Note that this can cause systems to trip the
kernel OOM killer on Linux systems if not enough physical memory and swap is not available.
The MAP_POPULATE option is used to populate pages into memory on systems that support this.
By default, anonymous mappings are used, however, the --mmap-file and --mmap-async options
allow one to perform file based mappings if desired.
Note that since stress-ng 0.17.05 the --mmap-madvise, --mmap-mergeable, --mmap-mprotect,
--mmap-slow-munmap and --mmap-write-check options should be used to enable the pre-0.17.05
mmap stressor behaviour.
--mmap-async
enable file based memory mapping and use asynchronous msync'ing on each page, see
--mmap-file.
--mmap-bytes N
allocate N bytes in total for all the mmap stressor instances, the default is 256 MB. One can
specify the size as % of total available memory or in units of Bytes, KBytes, MBytes and
GBytes using the suffix b, k, m or g.
--mmap-file
enable file based memory mapping and by default use synchronous msync'ing on each page.
--mmap-madvise
enable randomized madvise(2) settings on pages.
--mmap-mergeable
mark pages as mergeable via madvise(2) where possible.
--mmap-mlock
attempt to mlock mmap'd pages into memory causing more memory pressure by preventing pages
from swapped out.
--mmap-mmap2
use mmap2 for 4 K page aligned offsets if mmap2 is available, otherwise fall back to mmap.
--mmap-mprotect
change protection settings on each page of memory. Each time a page or a group of pages are
mapped or remapped then this option will make the pages read-only, write-only, exec-only, and
read-write.
--mmap-numa
assign memory mapped pages to randomly selected NUMA nodes. This is disabled for systems that
do not support NUMA.
--mmap-odirect
enable file based memory mapping and use O_DIRECT direct I/O.
--mmap-ops N
stop mmap stress workers after N bogo operations.
--mmap-osync
enable file based memory mapping and used O_SYNC synchronous I/O integrity completion.
--mmap-slow-munmap
enable page-by-page memory unmapping rather than attempting to memory unmap contiguous pages
in one large unmapping. This can cause lock contention when running with many concurrent mmap
stressors and will slow down the stressor.
--mmap-stressful
enable --mmap-file, --mmap-madvise, --mmap-mergeable, --mmap-mlock, --mmap-mprotect,
--mmap-odirect, --mmap-slow-munmap
--mmap-write-check
write into each page a unique 64 bit check value for all pages and then read the value for a
sanity check. This will force newly memory mapped pages to be faulted-in which slows down
mmap bogo-op rate. This can also cause lock contention on page allocation and page unmapping
on systems with many CPU threads and with cgroup memory accounting.
Random memory map/unmap stressor
--mmapaddr N
start N workers that memory map pages at a random memory location that is not already mapped.
On 64 bit machines the random address is randomly chosen 32 bit or 64 bit address. If the
mapping works a second page is memory mapped from the first mapped address. The stressor
exercises mmap/munmap, mincore and segfault handling.
--mmapaddr-mlock
attempt to mlock mmap'd pages into memory causing more memory pressure by preventing pages
from swapped out.
--mmapaddr-ops N
stop after N random address mmap bogo operations.
Memory map Copy-on-Write and unmap stressor
--mmapcow N
start N workers that exercise page allocation, modification (forcing copy-on-write) and page
unmapping. Regions of memory from 1 page to the maximum mappable size are allocated. Each
page in the regions are then modified, cache flushed (where possible) and unmapped. This
stressor produces a high page fault rate. For each region mapped there are 6 different
randomized strategies taken to exercise each page:
Sequential page modify and unmap
Sequential page modify and unmap (even pages, then odd pages)
Prime size page strides, page modify and unmap
Reverse sequential page modify and unmap
Reverse sequential page modify and unmap (even pages, then odd pages)
Single random page modify and unmap all pages
Randomly selected page modify and modify
--mmapcow-fork
fork child process to exercise pages in parallel with parent to exercice page handling
harder.
--mmapcow-free
after each page is modified and before each page is unmapped use madvise MADV_FREE (if it is
available) to indicate the page is free
--mmapcow-mlock
attempt to mlock each modified page, this can increase the number of page faults. Only
available with the mlock2 system call.
--mmapcow-numa
assign memory mapped pages to randomly selected NUMA nodes. This is disabled for systems that
do not support NUMA.
--mmapcow-ops N
stop after N pages are modified and unmapped
Forked memory map stressor
--mmapfork N
start N workers that each fork off 32 child processes, each of which tries to allocate some
of the free memory left in the system (and trying to avoid any swapping). The child
processes then hint that the allocation will be needed with madvise(2) and then memset it to
zero and hint that it is no longer needed with madvise before exiting. This produces
significant amounts of VM activity, a lot of cache misses and with minimal swapping.
--mmapfork-bytes N
specify the size of memory mapped fork region size. One can specify the size in units of
Bytes, KBytes, MBytes and GBytes using the suffix b, k, m or g.
--mmapfork-ops N
stop after N mmapfork bogo operations.
Memory map files stressor
--mmapfiles N
start N workers that attempt to memory map and then unmap up to 512 × 1024 files into memory.
The stressor will traverse /lib, /lib32, /lib64, /boot, /bin, /etc, /sbin, /usr, /var, /sys
and /proc and attempt to memory map files in these directories. Note that mapping bogo-ops
rate will depend on the speed of access to files on these file systems.
--mmapfiles-numa
assign memory mapped pages to randomly selected NUMA nodes. This is disabled for systems that
do not support NUMA.
--mmapfiles-ops N
stop after N memory map/unmap operations.
--mmapfiles-populate
The default is to perform a memory mapping and not fault any pages into physical memory. This
option uses MAP_POPULATE when available and will also read the first byte in each page to
ensure pages are faulted into memory to force memory population from file.
--mmapfiles-shared
The default is for private memory mapped files, however, with this option will use shared
memory mappings.
Fixed address memory map stressor
--mmapfixed N
start N workers that perform fixed address allocations from the top virtual address down to
128 K. The allocated sizes are from 1 page to 8 pages and various random mmap flags are used
MAP_SHARED/MAP_PRIVATE, MAP_LOCKED, MAP_NORESERVE, MAP_POPULATE. If successfully map'd then
the allocation is remap'd first to a large range of addresses based on a random start and
finally an address that is several pages higher in memory. Mappings and remappings are
madvised with random madvise options to further exercise the mappings.
--mmapfixed-mlock
attempt to mlock mmap'd pages into memory causing more memory pressure by preventing pages
from swapped out.
--mmapfixed-numa
assign memory mapped pages to randomly selected NUMA nodes. This is disabled for systems that
do not support NUMA.
--mmapfixed-ops N
stop after N mmapfixed memory mapping bogo operations.
Huge page memory mapping stressor
--mmaphuge N
start N workers that attempt to mmap a set of huge pages and large huge page sized mappings.
Successful mappings are madvised with MADV_NOHUGEPAGE and MADV_HUGEPAGE settings and then
1/64th of the normal small page size pages are touched. Finally, an attempt to unmap a small
page size page at the end of the mapping is made (these may fail on huge pages) before the
set of pages are unmapped. By default 8192 mappings are attempted per round of mappings or
until swapping is detected.
--mmaphuge-file
attempt to mmap on a 16 MB temporary file and random 4 K offsets. If this fails, anonymous
mappings are used instead.
--mmaphuge-mlock
attempt to mlock mmap'd huge pages into memory causing more memory pressure by preventing
pages from swapped out.
--mmaphuge-mmaps N
set the number of huge page mappings to attempt in each round of mappings. The default is
8192 mappings.
--mmaphuge-numa
assign memory mapped pages to randomly selected NUMA nodes. This is disabled for systems that
do not support NUMA.
--mmaphuge-ops N
stop after N mmaphuge bogo operations
Maximum memory mapping per process stressor
--mmapmany N
start N workers that attempt to create the maximum allowed per-process memory mappings. This
is achieved by mapping 3 contiguous pages and then unmapping the middle page hence splitting
the mapping into two. This is then repeated until the maximum allowed mappings or a maximum
of 262144 mappings are made.
--mmapmany-mlock
attempt to mlock mmap'd huge pages into memory causing more memory pressure by preventing
pages from swapped out.
--mmapmany-numa
assign memory mapped pages to randomly selected NUMA nodes. This is disabled for systems that
do not support NUMA.
--mmapmany-ops N
stop after N mmapmany bogo operations
Memory map torture stressor
--mmaptorture N
start N workers that exercise memory mapping operations on a shared file. Mappings of
various random sizes are tortured with calls to madvise, mremap, mprotect, msync, mlock,
munlock, mseal and mincore while the underlying file is modified by writes, truncation and
hole punching. Mapped data is read and written and sync'd back to file. For NUMA systems,
pages are randomly placed across NUMA nodes. All this activity causes cache misses, exercises
the TLB and causes IPI interrupts between CPUs.
--mmaptorture-bytes N
allocate N bytes in total for all the mmaptorture stressor instances, the default is 256 MB.
One can specify the size as % of total available memory or in units of Bytes, KBytes, MBytes
and GBytes using the suffix b, k, m or g.
--mmaptorture-msync N
memory mapped pages are msync'd various times per bogo-operation, this option specifies the
percentage amount of the pages per msync operation are msync'd. The default is 10%.
Specifying zero will disabled msync'ing.
--mmaptorture-ops N
stop after N iterations of memory mapping torture operations.
Kernel module loading stressor (Linux)
--module N
start N workers that use finit_module() to load the module specified or the hello test
module, if is available. There are different ways to test loading modules. Using modprobe
calls in a loop, using the kernel kernel module autoloader, and this stress-ng module
stressor. To stress tests modprobe we can simply run the userspace modprobe program in a
loop. To stress test the kernel module autoloader we can stress tests using the upstream
kernel tools/testing/selftests/kmod/kmod.sh. This ends up calling modprobe in the end, and it
has its own caps built-in to self protect the kernel from too many requests at the same time.
The userspace modprobe call will also prevent calls if the same module exists already. The
stress-ng modules stressor is designed to help stress test the finit_module() system call
even if the module is already loaded, testing races that are otherwise hard to reproduce.
--module-name NAME
NAME of the module to use, for example: test_module, xfs, ext4. By default test_module is
used so CONFIG_TEST_LKM must be enabled in the kernel. The module dependencies must be
loaded prior to running these stressor tests, as this stresses running finit_module() not
using modprobe.
--module-no-modver
ignore module modversions when using finit_module().
--module-no-vermag
ignore module versions when using finit_module().
--module-no-unload
do not unload the module right after loading it with finit_module().
--module-ops N
stop after N module load/unload cycles
Monte Carlo computations of π and e and various integrals
--monte-carlo N
start N stressors that compute π and e (Euler's number) using Monte Carlo computational
experiments with various random number generators.
--monte-carlo-method [ all | e | exp | pi | sin | sqrt | squircle ]
specify the computation to perform, options are as follows:
Method Description
all use all monte carlo computation methods
e compute Euler's constant e
exp integrate exp(x ↑ 2) for x = 0..1
pi compute π from the area of a circle
sin integrate sin(x) for x = 0..π
sqrt integrate sqrt(1 + x ↑ 4) for x = 0..1
squircle area of a unit squircle x ↑ 4 + y ↑ 4 = 1
--monte-carlo-ops N
stop after Monte Carlo computation experiments
--monte-carlo-rand [ all | drand48 | getrandom | lcg | pcg32 | mwc64 | random | xorshift ]
specify the random number generator to use, options are as follows:
Method Description
all use all the random number generators
arc4 use the libc cryptographically-secure pseudorandom arc4random(3) number generator.
drand48 use the libc linear congruential algorithm drand48(3) using 48-bit integer
arithmetic.
getrandom use the getrandom(2) system call for random values.
lcg use a 32 bit Paker-Miller Linear Congruential Generator, with a division
optimization.
pcg32 use a 32 bit O'Neill Permuted Congruential Generator.
mwc64 use the 64 bit stress-ng Multiply With Carry random number generator.
random use the libc random(3) Non-linear Additive Feedback random number generator.
xorshift use a 32 bit Marsaglia shift-register random number generator.
--monte-carlo-samples N
specify the number of random number samples to use to compute π or e, default is 100000.
Multi-precision floating operations (mpfr) stressor
--mpfr N
start N workers that exercise multi-precision floating point operations using the GNU Multi-
Precision Floating Point Reliable library (mpfr). Operations computed are as follows:
Method Description
apery calculate Apery's constant ζ(3); the sum of 1/(n ↑ 3).
cosine compute cos(θ) for θ = 0 to 2π in 100 steps.
euler compute e using n = (1 + (1 ÷ n)) ↑ n.
exp compute 1000 exponentials.
log computer 1000 natural logarithms.
omega compute the omega constant defined by Ωe↑Ω = 1 using efficient iteration of Ωn+1 = (1
+ Ωn) / (1 + e↑Ωn).
phi compute the Golden Ratio ϕ using series.
sine compute sin(θ) for θ = 0 to 2π in 100 steps.
nsqrt compute square root using Newton-Raphson.
--mpfr-ops N
stop workers after N iterations of various multi-precision floating point operations.
--mpfr-precision N
specify the precision in binary digits of the floating point operations. The default is 1000
bits, the allowed range is 32 to 1000000 (very slow).
Memory protection stressor
--mprotect N
start N workers that exercise changing page protection settings and access memory after each
change. 8 processes per worker contend with each other changing page protection settings on a
shared memory region of just a few pages to cause TLB flushes. A read and write to the pages
can cause segmentation faults and these are handled by the stressor. All combinations of page
protection settings are exercised including invalid combinations.
--mprotect-ops N
stop after N mprotect calls.
POSIX message queue stressor (Linux)
--mq N start N sender and receiver processes that continually send and receive messages using POSIX
message queues. (Linux only).
--mq-ops N
stop after N bogo POSIX message send operations completed.
--mq-size N
specify size of POSIX message queue. The default size is 10 messages and most Linux systems
this is the maximum allowed size for normal users. If the given size is greater than the
allowed message queue size then a warning is issued and the maximum allowed size is used
instead.
Memory remap stressor (Linux)
--mremap N
start N workers continuously calling mmap(2), mremap(2) and munmap(2). The initial anonymous
mapping is a large chunk (size specified by --mremap-bytes) and then iteratively halved in
size by remapping all the way down to a page size and then back up to the original size.
This worker is only available for Linux.
--mremap-bytes N
initially allocate N bytes per remap stress worker, the default is 256 MB. One can specify
the size in units of Bytes, KBytes, MBytes and GBytes using the suffix b, k, m or g.
--mremap-mlock
attempt to mlock remap'd pages into memory causing more memory pressure by preventing pages
from swapped out.
--mreamp-numa
assign memory mapped pages to randomly selected NUMA nodes. This is disabled for systems that
do not support NUMA.
--mremap-ops N
stop mremap stress workers after N bogo operations.
Memory Sealing
--mseal N
start N memory sealing stressors that exercise madvise, mremap, mprotect and mseal operations
on two pages of of memory mapped anonymous private memory. Linux 6.10+ kernels only.
--mseal-ops N
stop after N msealed memory operations.
System V message IPC stressor
--msg N
start N sender and receiver processes that continually send and receive messages using System
V message IPC.
--msg-bytes N
specify the size of the message being sent and received. Range 4 to 8192 bytes, default is 4
bytes.
--msg-ops N
stop after N bogo message send operations completed.
--msg-types N
select the quality of message types (mtype) to use. By default, msgsnd sends messages with a
mtype of 1, this option allows one to send messages types in the range 1..N to exercise the
message queue receive ordering. This will also impact throughput performance.
Synchronize file with memory map (msync) stressor
--msync N
start N stressors that msync data from a file backed memory mapping from memory back to the
file and msync modified data from the file back to the mapped memory. This exercises the
msync(2) MS_SYNC and MS_INVALIDATE sync operations.
--msync-bytes N
allocate N bytes for the memory mapped file, the default is 256 MB. One can specify the size
as % of total available memory or in units of Bytes, KBytes, MBytes and GBytes using the
suffix b, k, m or g.
--msync-ops N
stop after N msync bogo operations completed.
Synchronize file with memory map (msync) coherency stressor
--msyncmany N
start N stressors that memory map up to 32768 pages on the same page of a temporary file,
change the first 32 bits in a page and msync the data back to the file. The other 32767
pages are examined to see if the 32 bit check value is msync'd back to these pages.
--msyncmany-ops N
stop after N msync calls in the msyncmany stressors are completed.
ISO C mtx (mutex) stressor
--mtx N
start N stressors that exercise ISO C mutex locking and unlocking.
--mtx-ops N
stop after N bogo mutex lock/unlock operations.
--mtx-procs N
By default 2 threads are used for locking/unlocking on a single mutex. This option allows the
default to be changed to 2 to 64 concurrent threads.
Unmapping shared non-executable memory stressor (Linux)
--munmap N
start N stressors that exercise unmapping of shared non-executable mapped regions of child
processes (Linux only). The unmappings map shared memory regions page by page with a prime
sized stride that creates many temporary mapping holes. One the unmappings are complete the
child will exit and a new one is started. Note that this may trigger segmentation faults in
the child process, these are handled where possible by forcing the child process to call
_exit(2).
--munmap-ops N
stop after N page unmappings.
Pthread mutex stressor
--mutex N
start N stressors that exercise pthread mutex locking and unlocking. If run with enough
privilege then the FIFO scheduler is used and a random priority between 0 and 80% of the
maximum FIFO priority level is selected for the locking operation. The minimum FIFO priority
level is selected for the critical mutex section and unlocking operation to exercise random
inverted priority scheduling.
--mutex-affinity
enable random CPU affinity changing between mutex lock and unlock.
--mutex-ops N
stop after N bogo mutex lock/unlock operations.
--mutex-procs N
By default 2 threads are used for locking/unlocking on a single mutex. This option allows the
default to be changed to 2 to 64 concurrent threads.
High resolution and scheduler stressor via nanosleep calls
--nanosleep N
start N workers that each run pthreads that call nanosleep with random delays from 1 to 2↑18
nanoseconds. This should exercise the high resolution timers and scheduler.
--nanosleep-method [ all | cstate | random | ns | us | ms ]
select the nanosleep sleep duration method. By default, cstate residency durations (if they
exist) and random durations are used. This option allows one to select one of the three
methods:
Method Description
all use cstate and random nanosecond durations.
cstate use cstate nanosecond durations. It is recommended to also use --nanosleep-threads 1
to exercise less conconcurrent nanosleeps to allow CPUs to drop into deep C states.
random use random nanosecond durations between 1 and 2↑18 nanoseconds.
ns use 1ns (nanosecond) nanosleeps
us use 1us (microsecond) nanosleeps
ms use 1ms (millisecond) nanosleeps
--nanosleep-ops N
stop the nanosleep stressor after N bogo nanosleep operations.
--nanosleep-threads N
specify the number of concurrent pthreads to run per stressor. The default is 8 and the
allowed range is 1 to 1024.
Network device ioctl stressor
--netdev N
start N workers that exercise various netdevice ioctl commands across all the available
network devices. The ioctls exercised by this stressor are as follows: SIOCGIFCONF,
SIOCGIFINDEX, SIOCGIFNAME, SIOCGIFFLAGS, SIOCGIFADDR, SIOCGIFNETMASK, SIOCGIFMETRIC,
SIOCGIFMTU, SIOCGIFHWADDR, SIOCGIFMAP and SIOCGIFTXQLEN. See netdevice(7) for more details of
these ioctl commands.
--netdev-ops N
stop after N netdev bogo operations completed.
Netlink proc stressor (Linux)
--netlink-proc N
start N workers that spawn child processes and monitor fork/exec/exit process events via the
proc netlink connector. Each event received is counted as a bogo op. This stressor can only
be run on Linux and requires CAP_NET_ADMIN capability.
--netlink-proc-ops N
stop the proc netlink connector stressors after N bogo ops.
Netlink task stressor (Linux)
--netlink-task N
start N workers that collect task statistics via the netlink taskstats interface. This
stressor can only be run on Linux and requires CAP_NET_ADMIN capability.
--netlink-task-ops N
stop the taskstats netlink connector stressors after N bogo ops.
Nice stressor
--nice N
start N cpu consuming workers that exercise the available nice levels. Each iteration forks
off a child process that runs through the all the nice levels running a busy loop for 0.1
seconds per level and then exits.
--nice-ops N
stop after N nice bogo nice loops
NO-OP CPU instruction stressor
--nop N
start N workers that consume cpu cycles issuing no-op instructions. This stressor is
available if the assembler supports the "nop" instruction.
--nop-instr INSTR
use alternative nop instruction INSTR. For x86 CPUs INSTR can be one of nop, pause, nop2 (2
byte nop) through to nop15 (15 byte nop) and fnop. For ARM CPUs, INSTR can be one of nop or
yield. For PPC64 CPUs, INSTR can be one of nop, mdoio, mdoom or yield. For S390 CPUs, INSTR
can be one of nop or nopr. For other processors, INSTR is only nop. The random INSTR option
selects a random mix of the available nop instructions. If the chosen INSTR generates an
SIGILL signal, then the stressor falls back to the vanilla nop instruction.
--nop-ops N
stop nop workers after N no-op bogo operations. Each bogo-operation is equivalent to 256
loops of 256 no-op instructions.
/dev/null stressor
--null N
start N workers that exercise /dev/null with writes, lseek, ioctl, fcntl, fallocate and
fdatasync. For just /dev/null write benchmarking use the --null-write option.
--null-ops N
stop null stress workers after N /dev/null bogo operations.
--null-write
just write to /dev/null with 4 K writes with no additional exercising on /dev/null.
Migrate memory pages over NUMA nodes stressor
--numa N
start N workers that migrate stressors and a 4 MB memory mapped buffer around all the
available NUMA nodes. This uses migrate_pages(2) to move the stressors and mbind(2) and
move_pages(2) to move the pages of the mapped buffer. After each move, the buffer is written
to force activity over the bus which results cache misses. This test will only run on
hardware with NUMA enabled.
--numa-bytes N
specify the total number bytes to be exercised by all the workers, the given size is divided
by the number of workers and rounded to the nearest page size. The default is 4 MB per
worker. One can specify the size as % of total available memory or in units of Bytes, KBytes,
MBytes and GBytes using the suffix b, k, m or g.
--numa-ops N
stop NUMA stress workers after N bogo NUMA operations.
--numa-shuffle-addr
shuffle page order for the address list when calling move_pages(2)
--numa-shuffle-node
shuffle node order for the address list when calling move_pages(2)
Large Pipe stressor
--oom-pipe N
start N workers that create as many pipes as allowed and exercise expanding and shrinking the
pipes from the largest pipe size down to a page size. Data is written into the pipes and read
out again to fill the pipe buffers. With the --aggressive mode enabled the data is not read
out when the pipes are shrunk, causing the kernel to OOM processes aggressively. Running
many instances of this stressor will force kernel to OOM processes due to the many large pipe
buffer allocations.
--oom-pipe-ops N
stop after N bogo pipe expand/shrink operations.
Illegal instructions stressors
--opcode N
start N workers that fork off children that execute randomly generated executable code. This
will generate issues such as illegal instructions, bus errors, segmentation faults, traps,
floating point errors that are handled gracefully by the stressor.
--opcode-method [ inc | mixed | random | text ]
select the opcode generation method. By default, random bytes are used to generate the
executable code. This option allows one to select one of the three methods:
Method Description
inc use incrementing 32 bit opcode patterns from 0x00000000 to 0xfffffff inclusive.
mixed use a mix of incrementing 32 bit opcode patterns and random 32 bit opcode patterns
that are also inverted, encoded with gray encoding and bit reversed.
random generate opcodes using random bytes from a mwc random generator.
text copies random chunks of code from the stress-ng text segment and randomly flips
single bits in a random choice of 1/8th of the code.
--opcode-ops N
stop after N attempts to execute illegal code.
Opening file (open) stressor
-o N, --open N
start N workers that perform open(2) and then close(2) operations on /dev/zero. The maximum
opens at one time is system defined, so the test will run up to this maximum, or 65536 open
file descriptors, which ever comes first.
--open-fd
run a child process that scans /proc/$PID/fd and attempts to open the files that the stressor
has opened. This exercises racing open/close operations on the proc interface.
--open-max N
try to open a maximum of N files (or up to the maximum per-process open file system limit).
The value can be the number of files or a percentage of the maximum per-process open file
system limit.
--open-ops N
stop the open stress workers after N bogo open operations.
Page table and TLB stressor
--pagemove N
start N workers that mmap a memory region (default 4 MB) and then shuffle pages to the
virtual address of the previous page. Each page shuffle uses 3 mremap operations to move a
page. This exercises page tables and Translation Lookaside Buffer (TLB) flushing.
--pagemove-bytes
specify the size of the memory mapped region to be exercised. One can specify the size as %
of total available memory or in units of Bytes, KBytes, MBytes and GBytes using the suffix b,
k, m or g.
--pagemove-mlock
attempt to mlock mmap'd and mremap'd pages into memory causing more memory pressure by
preventing pages from swapped out.
--pagemove-numa
assign memory mapped pages to randomly selected NUMA nodes. This is disabled for systems that
do not support NUMA.
--pagemove-ops N
stop after N pagemove shuffling operations, where suffling all the pages in the mmap'd region
is equivalent to 1 bogo-operation.
Memory page swapping stressor
--pageswap N
start N workers that exercise page swap in and swap out. Pages are allocated and paged out
using madvise MADV_PAGEOUT. One the maximum per process number of mmaps are reached or 65536
pages are allocated the pages are read to page them back in and unmapped in reverse mapping
order.
--pageswap-ops N
stop after N page allocation bogo operations.
PCI sysfs stressor (Linux)
--pci N
exercise PCI sysfs by running N workers that read data (and mmap/unmap PCI config or PCI
resource files). Linux only. Running as root will allow config and resource mmappings to be
read and exercises PCI I/O mapping.
--pci-dev xxxx:xx:xx.x
specify a PCI device to exercise rather than exercise all PCI devices. The device is
specified using the PCI device name xxxx:xx:xx.x where x is a hexadecimal digit.
--pci-ops N
stop pci stress workers after N PCI subdirectory exercising operations.
--pci-ops-rate N
specify the PCI bogo-ops per second rate. This is useful for PCI bandwidth limiting and on
x86 systems this may reduce uncore transactions.
Personality stressor
--personality N
start N workers that attempt to set personality and get all the available personality types
(process execution domain types) via the personality(2) system call. (Linux only).
--personality-ops N
stop personality stress workers after N bogo personality operations.
Mutex using Peterson algorithm stressor
--peterson N
start N workers that exercises mutex exclusion between two processes using shared memory with
the Peterson Algorithm. Where possible this uses memory fencing and falls back to using GCC
__sync_synchronize if they are not available. The stressors contain simple mutex and memory
coherency sanity checks.
--peterson-ops N
stop peterson workers after N mutex operations.
Page mmap stressor
--physpage N
start N workers that use /proc/self/pagemap and /proc/kpagecount to determine the physical
page and page count of a virtual mapped page and a page that is shared among all the
stressors. Linux only and requires the CAP_SYS_ADMIN capabilities.
--physpage-mtrr
enable setting various memory type rage register (MTRR) types on physical pages (Linux and
x86 only).
--physpage-ops N
stop physpage stress workers after N bogo physical address lookups.
Physical page mmap stressor (using /dev/mem)
--physmmap N
start N workers that try to mmap available System RAM (as specified by /proc/iomem) using
/dev/mem and randomized private or shared pages. This requires CAP_SYS_ADMIN capabilities to
fully read /proc/iomem and mmap onto /dev/mem. Note systems may be fully locked down and
mmapping onto /dev/mem is impossible.
--physmmap-ops N
stop physmmap stress workers after N bogo mmap attempts.
--physmmap-read
perform 64 bit reads of all the data in each mmap'd page.
Process signals (pidfd_send_signal) stressor
--pidfd N
start N workers that exercise signal sending via the pidfd_send_signal system call. This
stressor creates child processes and checks if they exist and can be stopped, restarted and
killed using the pidfd_send_signal system call.
--pidfd-ops N
stop pidfd stress workers after N child processes have been created, tested and killed with
pidfd_send_signal.
Localhost ICMP (ping) stressor
--ping-sock N
start N workers that send small randomized ICMP messages to the localhost across a range of
ports (1024..65535) using a "ping" socket with an AF_INET domain, a SOCK_DGRAM socket type
and an IPPROTO_ICMP protocol.
--ping-sock-ops N
stop the ping-sock stress workers after N ICMP messages are sent.
Large pipe stressor
-p N, --pipe N
start N workers that perform large pipe writes and reads to exercise pipe I/O. This
exercises memory write and reads as well as context switching. Each worker has two
processes, a reader and a writer.
--pipe-data-size N
specifies the size in bytes of each write to the pipe (range from 4 bytes to 4096 bytes).
Setting a small data size will cause more writes to be buffered in the pipe, hence reducing
the context switch rate between the pipe writer and pipe reader processes. Default size is
the page size.
--pipe-ops N
stop pipe stress workers after N bogo pipe write operations.
--pipe-vmsplice
use vmsplice(2) to splice data pages to/from pipe. Requires pipe packet mode using O_DIRECT
and buffer twice the size of the pipe to ensure verification data sequences.
--pipe-size N
specifies the size of the pipe in bytes (for systems that support the F_SETPIPE_SZ fcntl()
command). Setting a small pipe size will cause the pipe to fill and block more frequently,
hence increasing the context switch rate between the pipe writer and the pipe reader
processes. As of version 0.15.11 the default size is 4096 bytes.
Shared pipe stressor
--pipeherd N
start N workers that pass a 64 bit token counter to/from 100 child processes over a shared
pipe. This forces a high context switch rate and can trigger a "thundering herd" of wakeups
on processes that are blocked on pipe waits.
--pipeherd-ops N
stop pipe stress workers after N bogo pipe write operations.
--pipeherd-yield
force a scheduling yield after each write, this increases the context switch rate.
Memory protection key mechanism stressor (Linux)
--pkey N
start N workers that change memory protection using a protection key (pkey) and the
pkey_mprotect call (Linux only). This will try to allocate a pkey and use this for the page
protection, however, if this fails then the special pkey -1 will be used (and the kernel will
use the normal mprotect mechanism instead). Various page protection mixes of
read/write/exec/none will be cycled through on randomly chosen pre-allocated pages.
--pkey-ops N
stop after N pkey_mprotect page protection cycles.
Stress-ng plugin stressor
--plugin N
start N workers that run user provided stressor functions loaded from a shared library. The
shared library can contain one or more stressor functions prefixed with stress_ in their
name. By default the plugin stressor will find all functions prefixed with stress_ in their
name and exercise these one by one in a round-robin loop, but a specific stressor can be
selected using the --plugin-method option. The stressor function takes no parameters and
returns 0 for success and non-zero for failure (and will terminate the plugin stressor). Each
time a stressor function is executed the bogo-op counter is incremented by one. The following
example performs 10,000 nop instructions per bogo-op:
int stress_example(void)
{
int i;
for (i = 0; i < 10000; i++) {
__asm__ __volatile__("nop");
}
return 0; /* Success */
}
and compile the source into a shared library as, for example:
gcc -fpic -shared -o example.so example.c
and run as using:
stress-ng --plugin 1 --plugin-so ./example.so
--plugin-method function
run a specific stressor function, specify the name without the leading stress_ prefix.
--plugin-ops N
stop after N iterations of the user provided stressor function(s).
--plugin-so name
specify the shared library containing the user provided stressor function(s).
Polling stressor
-P N, --poll N
start N workers that perform zero timeout polling via the poll(2), ppoll(2), select(2),
pselect(2) and sleep(3) calls. This wastes system and user time doing nothing.
--poll-fds N
specify the number of file descriptors to poll/ppoll/select/pselect on. The maximum number
for select/pselect is limited by FD_SETSIZE and the upper maximum is also limited by the
maximum number of pipe open descriptors allowed.
--poll-ops N
stop poll stress workers after N bogo poll operations.
--poll-random-us N
use a random timeout of N microseconds for ppoll(2) and pselect(2) calls instead of the
default of 20000 microseconds.
Power maths functions
--powmath N
start N workers that exercise various power functions with input values 0 to 1 in steps of
0.001; the results are sanity checked to ensure no variation occurs after each round of 10000
computations.
--powmath-ops N
stop after N power function bogo-operation loops.
--powmath-method method
specify a power function to exercise. Available power function stress methods are described
as follows:
Method Description
all iterate over all the below power functions methods
cpow complex double power function
cpowf complex float power function
cpowl complex long double power function
csqrt complex double square root function (1/2 power)
csqrtf complex float square root function (1/2 power)
csqrtl complex long double square root function (1/2 power)
cbrt double cube root function (1/3 power)
cbrtf float cube root function (1/3 power)
cbrtl long double cube root function (1/3 power)
hypot double Euclidean distance function (hypotenuse)
hypotf float Euclidean distance function (hypotenuse)
hypotl long double Euclidean distance function (hypotenuse)
pow double power function
powf float power function
powl long double power function
sqrt double square root function (1/2 power)
sqrtf float square root function (1/2 power)
sqrtl long double square root function (1/2 power)
Prctl stressor
--prctl N
start N workers that exercise the majority of the prctl(2) system call options. Each batch of
prctl calls is performed inside a new child process to ensure the limit of prctl is contained
inside a new process every time. Some prctl options are architecture specific, however, this
stressor will exercise these even if they are not implemented.
--prctl-ops N
stop prctl workers after N batches of prctl calls
L3 cache prefetching stressor
--prefetch N
start N workers that benchmark prefetch and non-prefetch reads of a L3 cache sized buffer.
The buffer is read with loops of 8 × 64 bit reads per iteration. In the prefetch cases, data
is prefetched ahead of the current read position by various sized offsets, from 64 bytes to 8
K to find the best memory read throughput. The stressor reports the non-prefetch read rate
and the best prefetched read rate. It also reports the prefetch offset and an estimate of the
amount of time between the prefetch issue and the actual memory read operation. These
statistics will vary from run-to-run due to system noise and CPU frequency scaling.
--prefetch-l3-size N
specify the size of the l3 cache
--prefetch-method N
select the prefetching method. Available methods are:
Method Description
builtin Use the __builtin_prefetch(3) function for prefetching. This is the default.
builtinl0 Use the __builtin_prefetch(3) function for prefetching, with a locality 0
hint.
builtinl3 Use the __builtin_prefetch(3) function for prefetching, with a locality 3
hint.
dcbt Use the ppc64 dcbt instruction to fetch data into the L1 cache (ppc64 only).
dcbtst Use the ppc64 dcbtst instruction to fetch data into the L1 cache (ppc64
only).
prefetcht0 Use the x86 prefetcht0 instruction to prefetch data into all levels of the
cache hierarchy (x86 only).
prefetcht1 Use the x86 prefetcht1 instruction (temporal data with respect to first level
cache) to prefetch data into level 2 cache and higher (x86 only).
prefetcht2 Use the x86 prefetcht2 instruction (temporal data with respect to second
level cache) to prefetch data into level 2 cache and higher (x86 only).
prefetchnta Use the x86 prefetchnta instruction (non-temporal data with respect to all
cache levels) into a location close to the processor, minimizing cache
pollution (x86 only).
prfm_pldl1keep Use the aarch64 prfm instruction with pld1keep operation (retained or
temporal prefetch, allocated in the cache normally) to prefetch data into
level 1 of the cache hierarchy (aarch64 only).
prfm_pldl2keep Use the aarch64 prfm instruction with pld1keep operation (retained or
temporal prefetch, allocated in the cache normally) to prefetch data into
level 2 of the cache hierarchy (aarch64 only).
prfm_pldl3keep Use the aarch64 prfm instruction with pld1keep operation (retained or
temporal prefetch, allocated in the cache normally) to prefetch data into
level 3 of the cache hierarchy (aarch64 only).
prfm_pldl1strm Use the aarch64 prfm instruction with pld1strm operation (streaming or non-
temporal prefetch, for data that is used only once) to prefetch data into
level 1 of the cache hierarchy (aarch64 only).
prfm_pldl2strm Use the aarch64 prfm instruction with pld1strm operation (streaming or non-
temporal prefetch, for data that is used only once) to prefetch data into
level 2 of the cache hierarchy (aarch64 only).
prfm_pldl3strm Use the aarch64 prfm instruction with pld1strm operation (streaming or non-
temporal prefetch, for data that is used only once) to prefetch data into
level 3 of the cache hierarchy (aarch64 only).
--prefetch-ops N
stop prefetch stressors after N benchmark operations
Search for prime numbers using large integers
--prime N
start N workers that find prime numbers using the GNU Multiple Precision Arithmetic Library
for large integers. The GMP mpz_nextprime function is used to find primes and it uses a
probabilistic algorithm to identify primes, but there is a extremely small chance that the
values found are non-prime. The search becomes computationally more expensive over time to
find larger and larger primes, hence the bogo-op rate will reduce over time.
--prime-method [ factorial | inc | pwr2 | pwr10 ]
selects the method of calculating the next value from where to start searching primes and
hence how large the primes get. The default is inc, the methods to start searching for primes
are described as follows.
Method Description
factorial start of search based on factorial expansion. This grows rapidly.
inc start of search based on increments by 2. This grows very slowly.
pwr2 start of search based on powers of 2. Grows relatively quickly.
pwr10 start of search based on powers of 10. Grows by 1 digit per iteration and grows
quickly.
--prime-ops N
stop after finding N prime numbers.
--prime-progress
show the number of primes found and length of largest prime found. This is displayed either
every 60 seconds or more than 60 seconds if it takes longer to find the next prime.
--prime-start N
start the prime search from value N. The value may be expressed as an integer value or as a
floating point value (e.g. 1e200 to express a very large starting value).
Priority inversion stressor
--prio-inv N
start N workers that exercise mutex lock priority inversion scheduling. Three child process
run with low, medium and high FIFO scheduling priorities. The processes with low and high
priorities share a mutex lock that both try to lock and unlock, aiming to make the low
priority process block the high priority process. Meanwhile the middle priority process will
run in priority over the low priority process, causing the high priority process to become
unrunnable.
--prio-inv-ops N
stop after N bogo lock/unlock operations.
--prio-inv-type [ inherit | none | protect ]
select the mutex lock priority inversion type, described as follows:
Type Description
inherit The priority of the process owning the mutex lock is run with highest priority of
any other process waiting on the lock to avoid priority inversion deadlock.
none The priority of the process owning the mutex lock is not affected by its mutex
ownership. This may lead to the high priority process to become unrunnable on a
single thread system.
protect The priority of the process owning the mutex lock is given the priority of the mutex
(in this stress test case, the maximum priority) during the lock ownership.
--prio-inv-policy [ batch | idle | fifo | other | rr ]
select the scheduling policy. "Normal" policies (batch, idle and other) can be selected as an
unprivileged user, however "Real Time" policies (fifo and rr) can only be selected with the
appropriate privilege. By default "fifo" is selected but it will fall back to "other" for
unprivileged users.
Privileged CPU instructions stressor
--priv-instr N
start N workers that exercise various architecture specific privileged instructions that
cannot be executed by userspace programs. These instructions will be trapped and processed by
SIGSEGV or SIGILL signal handlers.
--priv-instr-ops N
stop priv-instr stressors after N rounds of executing privileged instructions.
/proc stressor
--procfs N
start N workers that read files from /proc and recursively read files from /proc/self (Linux
only).
--procfs-ops N
stop procfs reading after N bogo read operations. Note, since the number of entries may vary
between kernels, this bogo ops metric is probably very misleading.
pwrite/pread lseek I/O stressor
--pseek N
start N workers that exercise pwrite() and pread() with lseek() positioning tests. Each
worker has 4 sub-processes that perform repeated pwrite() and pread() operations using the
same shared file descriptor. Two of the processes are started using pthreads (if available),
another two processes are started with fork(). Using pwrite() and pread() should perform I/O
without altering the shared file descriptor file offset. The main worker process performs I/O
using lseek()/write() and lseek()/read() calls that change the file offset; lseeks are sanity
checked to see if they are being altered by pwrite() and pread() calls.
--pseek-io-size N
specify size of each write/read I/O operation in bytes. Size can be from 1 byte to 1 MB.
--pseek-ops N
stop after N writes by the main worker process.
--pseek-rand
normally each sub-process writes to a fix file offset, using this option will randomize the
offset on each write/read cycle.
Pthread stressor
--pthread N
start N workers that iteratively creates and terminates multiple pthreads (the default is
1024 pthreads per worker). In each iteration, each newly created pthread waits until the
worker has created all the pthreads and then they all terminate together.
--pthread-max N
create N pthreads per worker. If the product of the number of pthreads by the number of
workers is greater than the soft limit of allowed pthreads then the maximum is re-adjusted
down to the maximum allowed.
--pthread-ops N
stop pthread workers after N bogo pthread create operations.
Ptrace stressor
--ptrace N
start N workers that fork and trace system calls of a child process using ptrace(2).
--ptrace-ops N
stop ptracer workers after N bogo system calls are traced.
Pointer Chasing stressor
--ptr-chase N
start N workers that chase memory pointers around page sized nodes of pointers. By default
each stressor instance allocates 4096 pages of nodes, each node is an array of pointers that
are randomly set to point to other node pages. The stressors follow randomly chosen pointers
from each node, chasing around the entire node space. This exercises pointer fetching,
pointer dereferencing and is a cache-read exercising stressor. The nodes are allocated with
50% of pages from the heap and 50% from mmap'd memory.
--ptr-chase-ops N
stop after N pointer chases
--ptr-chase-pages N
select number of pages to allocate for the nodes.
Pseudo-terminals (pty) stressor
--pty N
start N workers that repeatedly attempt to open pseudoterminals and perform various pty
ioctls upon the ptys before closing them.
--pty-max N
try to open a maximum of N pseudoterminals, the default is 65536. The allowed range of this
setting is 8..65536.
--pty-ops N
stop pty workers after N pty bogo operations.
Qsort stressor
-Q, --qsort N
start N workers that sort 32 bit integers using qsort.
--qsort-method [ qsort-libc | qsort-bm ]
select either the libc implementation of qsort or the J. L. Bentley and M. D. McIlroy
implementation of qsort. The default is the libc implementation.
--qsort-ops N
stop qsort stress workers after N bogo qsorts.
--qsort-size N
specify number of 32 bit integers to sort, default is 262144 (256 × 1024).
Quota stressor
--quota N
start N workers that exercise the Q_GETQUOTA, Q_GETFMT, Q_GETINFO, Q_GETSTATS and Q_SYNC
quotactl(2) commands on all the available mounted block based file systems. Requires
CAP_SYS_ADMIN capability to run.
--quota-ops N
stop quota stress workers after N bogo quotactl operations.
Process scheduler stressor
--race-sched N
start N workers that exercise rapid changing CPU affinity child processes both from the
controlling stressor and by the child processes. Child processes are created and terminated
rapidly with the aim to create race conditions where affinity changing occurs during process
run states.
--race-sched-method [ all | next | prev | rand | randinc | syncnext | syncprev ]
Select the method moving a process to a specific CPU. Available methods are described as
follows:
Method Description
all iterate over all the race-sched methods as listed below:
next move a process to the next CPU, wrap around to zero when maximum CPU is reached.
prev move a process to the previous CPU, wrap around to the maximum CPU when the first
CPU is reached.
rand move a process to any randomly chosen CPU.
randinc move a process to the current CPU + a randomly chosen value 1..4, modulo the number
of CPUs.
syncnext move synchronously all the race-sched stressor processes to the next CPU every
second; this loads just 1 CPU at a time in a round-robin method.
syncprev move synchronously all the race-sched stressor processes to the previous CPU every
second; this loads just 1 CPU at a time in a round-robin method.
--race-sched-ops N
stop after N process creation bogo-operations.
Radixsort stressor
--radixsort N
start N workers that sort random 8 byte strings using radixsort.
--radixsort-method [ radixsort-libc | radixsort-nonlibc ]
select either the libc implementation of radixsort or an optimized implementation of
radixsort. The default is the libc implementation if it is available.
--radixsort-ops N
stop radixsort stress workers after N bogo radixsorts.
--radixsort-size N
specify number of strings to sort, default is 262144 (256 × 1024).
Memory filesystem stressor
--ramfs N
start N workers mounting a memory based file system using ramfs and tmpfs (Linux only). This
alternates between mounting and umounting a ramfs or tmpfs file system using the traditional
mount(2) and umount(2) system call as well as the newer Linux 5.2 fsopen(2), fsmount(2),
fsconfig(2) and move_mount(2) system calls if they are available. The default ram file system
size is 2 MB.
--ramfs-fill
fill ramfs with zero'd data using fallocate(2) if it is available or multiple calls to
write(2) if not.
--ramfs-ops N
stop after N ramfs mount operations.
--ramfs-size N
set the ramfs size (must be multiples of the page size).
Raw device stressor
--rawdev N
start N workers that read the underlying raw drive device using direct IO reads. The device
(with minor number 0) that stores the current working directory is the raw device to be read
by the stressor. The read size is exactly the size of the underlying device block size. By
default, this stressor will exercise all the of the rawdev methods (see the --rawdev-method
option). This is a Linux only stressor and requires root privilege to be able to read the raw
device.
--rawdev-method method
Available rawdev stress methods are described as follows:
Method Description
all iterate over all the rawdev stress methods as listed below:
sweep repeatedly read across the raw device from the 0th block to the end block in steps of
the number of blocks on the device / 128 and back to the start again.
wiggle repeatedly read across the raw device in 128 evenly steps with each step reading 1024
blocks backwards from each step.
ends repeatedly read the first and last 128 start and end blocks of the raw device
alternating from start of the device to the end of the device.
random repeatedly read 256 random blocks
burst repeatedly read 256 sequential blocks starting from a random block on the raw device.
--rawdev-ops N
stop the rawdev stress workers after N raw device read bogo operations.
Random list stressor
--randlist N
start N workers that creates a list of objects in randomized memory order and traverses the
list setting and reading the objects. This is designed to exerise memory and cache thrashing.
Normally the objects are allocated on the heap, however for objects of page size or larger
there is a 1 in 16 chance of objects being allocated using shared anonymous memory mapping to
mix up the address spaces of the allocations to create more TLB thrashing.
--randist-compact
Allocate all the list objects using one large heap allocation and divide this up for all the
list objects. This removes the overhead of the heap keeping track of each list object, hence
uses less memory.
--randlist-items N
Allocate N items on the list. By default, 100,000 items are allocated.
--randlist-ops N
stop randlist workers after N list traversals
--randlist-size N
Allocate each item to be N bytes in size. By default, the size is 64 bytes of data payload
plus the list handling pointer overhead.
Localhost raw socket stressor
--rawsock N
start N workers that send and receive packet data using raw sockets on the localhost.
Requires CAP_NET_RAW to run.
--rawsock-ops N
stop rawsock workers after N packets are received.
--rawsock-port P
start at socket port P. For N rawsock worker processes, ports P to P − 1 are used.
Localhost ethernet raw packets stressor
--rawpkt N
start N workers that sends and receives ethernet packets using raw packets on the localhost
via the loopback device. Requires CAP_NET_RAW to run.
--rawpkt-ops N
stop rawpkt workers after N packets from the sender process are received.
--rawpkt-port N
start at port P. For N rawpkt worker processes, ports P to (P × 4) − 1 are used. The default
starting port is port 14000.
--rawpkt-rxring N
setup raw packets with RX ring with N number of blocks, this selects TPACKET_V. N must be one
of 1, 2, 4, 8 or 16.
Localhost raw UDP packet stressor
--rawudp N
start N workers that send and receive UDP packets using raw sockets on the localhost.
Requires CAP_NET_RAW to run.
--rawudp-if NAME
use network interface NAME. If the interface NAME does not exist, is not up or does not
support the domain then the loopback (lo) interface is used as the default.
--rawudp-ops N
stop rawudp workers after N packets are received.
--rawudp-port N
start at port P. For N rawudp worker processes, ports P to (P × 4) - 1 are used. The default
starting port is port 13000.
Random number generator stressor
--rdrand N
start N workers that read a random number from an on-chip random number generator This uses
the rdrand instruction on Intel x86 processors or the darn instruction on Power9 processors.
--rdrand-ops N
stop rdrand stress workers after N bogo rdrand operations (1 bogo op = 2048 random bits
successfully read).
--rdrand-seed
use rdseed instead of rdrand (x86 only).
Read-ahead stressor
--readahead N
start N workers that randomly seek and perform 4096 byte read/write I/O operations on a file
with readahead. The default file size is 64 MB. Readaheads and reads are batched into 16
readaheads and then 16 reads.
--readahead-bytes N
set the size of readahead file, the default is 1 GB. One can specify the size as % of free
space on the file system or in units of Bytes, KBytes, MBytes and GBytes using the suffix b,
k, m or g.
--readahead-ops N
stop readahead stress workers after N bogo read operations.
Reboot stressor
--reboot N
start N workers that exercise the reboot(2) system call. When possible, it will create a
process in a PID namespace and perform a reboot power off command that should shutdown the
process. Also, the stressor exercises invalid reboot magic values and invalid reboots when
there are insufficient privileges that will not actually reboot the system.
--reboot-ops N
stop the reboot stress workers after N bogo reboot cycles.
POSIX regular expressions stressor
--regex N
start N workers that compile various POSIX regular expressions and execute them against a set
of text strings. This exercises the regex C library with a range of various simple and
complex regex expressions.
--regex-ops N
stop after N regex compilations.
CPU registers stressor
--regs N
start N workers that shuffle data around the CPU registers exercising register move
instructions. Each bogo-op represents 1000 calls of a shuffling function that shuffles the
registers 32 times. Only implemented for the GCC compiler since this requires register
annotations and optimization level 0 to compile appropriately.
--regs-ops N
stop regs stressors after N bogo operations.
Memory page reordering stressor
--remap N
start N workers that map 512 pages and re-order these pages using the deprecated system call
remap_file_pages(2). Several page re-orderings are exercised: forward, reverse, random and
many pages to 1 page.
--remap-mlock
attempt to mlock mmap'd huge pages into memory causing more memory pressure by preventing
pages from swapped out.
--remap-ops N
stop after N remapping bogo operations.
--remap-pages N
specify number of pages to remap, must be a power of 2, default is 512 pages.
Renaming file stressor
-R N, --rename N
start N workers that each create a file and then repeatedly rename it.
--rename-ops N
stop rename stress workers after N bogo rename operations.
Process rescheduling stressor
--resched N
start N workers that exercise process rescheduling. Each stressor spawns a child process for
each of the positive nice levels and iterates over the nice levels from 0 to the lowest
priority level (highest nice value). For each of the nice levels 1024 iterations over 3 non-
real time scheduling polices SCHED_OTHER, SCHED_BATCH and SCHED_IDLE are set and a
sched_yield occurs to force heavy rescheduling activity. When the -v verbose option is used
the distribution of the number of yields across the nice levels is printed for the first
stressor out of the N stressors.
--resched-ops N
stop after N rescheduling sched_yield calls.
System resources stressor
--resources N
start N workers that consume various system resources. Each worker will spawn 1024 child
processes that iterate 1024 times consuming shared memory, heap, stack, temporary files and
various file descriptors (eventfds, memoryfds, userfaultfds, pipes and sockets).
--resources-mlock
attempt to mlock mmap'd pages into memory causing more memory pressure by preventing pages
from swapped out.
--resources-ops N
stop after N resource child forks.
Writing temporary files in reverse position stressor
--revio N
start N workers continually writing in reverse position order to temporary files. The default
mode is to stress test reverse position ordered writes with randomly sized sparse holes
between each write. With the --aggressive option enabled without any --revio-opts options
the revio stressor will work through all the --revio-opt options one by one to cover a range
of I/O options.
--revio-bytes N
write N bytes for each revio process, the default is 1 GB. One can specify the size as % of
free space on the file system or in units of Bytes, KBytes, MBytes and GBytes using the
suffix b, k, m or g.
--revio-opts list
specify various stress test options as a comma separated list. Options are the same as
--hdd-opts but without the iovec option and with a readahead option to force a readahead
action over the entire file once it has been completely written.
--revio-ops N
stop revio stress workers after N bogo operations.
--revio-write-size N
specify size of each write in bytes. Size can be from 1 byte to 4 MB.
Ring pipes stressor
--ring-pipe N
start N workers that move data around a ring of pipes using poll to detect when data is ready
to copy. By default, 256 pipes are used with two 4096 byte items of data being copied around
the ring of pipes. Data is copied using read and write system calls. If the splice system
call is available then one can use splice to use more efficient in-kernel data passing
instead of buffer copying.
--ring-pipe-num N
specify the number of pipes to use. Ranges from 4 to 262144, default is 256.
--ring-pipe-ops N
stop after N pipe data transfers.
--ring-pipe-size N
specify the size of data being copied in bytes. Ranges from 1 to 4096, default is 4096.
--ring-pipe-splice
enable splice to move data between pipes (only if splice() is available).
Rlimit stressor
--rlimit N
start N workers that exceed CPU and file size resource imits, generating SIGXCPU and SIGXFSZ
signals.
--rlimit-ops N
stop after N bogo resource limited SIGXCPU and SIGXFSZ signals have been caught.
VM reverse-mapping stressor
--rmap N
start N workers that exercise the VM reverse-mapping. This creates 16 processes per worker
that write/read multiple file-backed memory mappings. There are 64 lots of 4 page mappings
made onto the file, with each mapping overlapping the previous by 3 pages and at least 1 page
of non-mapped memory between each of the mappings. Data is synchronously msync'd to the file
1 in every 256 iterations in a random manner.
--rmap-ops N
stop after N bogo rmap memory writes/reads.
1 bit rotation stressor
--rotate N
start N workers that exercise 1 bit rotates left and right of unsigned integer variables.
The default will rotate four 8, 16, 32, 64 (and if supported 128) bit values 10000 times in a
loop per bogo-op.
--rotate-method method
specify the method of rotation to use. The `all' method uses all the methods and is the
default.
Method Description
all exercise with all the rotate stressor methods (see below):
rol8 8 bit unsigned rotate left by 1 bit
ror8 8 bit unsigned rotate right by 1 bit
rol16 16 bit unsigned rotate left by 1 bit
ror16 16 bit unsigned rotate right by 1 bit
rol32 32 bit unsigned rotate left by 1 bit
ror32 32 bit unsigned rotate right by 1 bit
rol64 64 bit unsigned rotate left by 1 bit
ror64 64 bit unsigned rotate right by 1 bit
rol128 128 bit unsigned rotate left by 1 bit
ror128 128 bit unsigned rotate right by 1 bit
--rotate-ops N
stop after N bogo rotate operations.
Restartable sequences (rseq) stressor (Linux)
--rseq N
start N workers that exercise restartable sequences via the rseq(2) system call. This loops
over a long duration critical section that is likely to be interrupted. A rseq abort handler
keeps count of the number of interruptions and a SIGSEV handler also tracks any failed rseq
aborts that can occur if there is a mismatch in a rseq check signature. Linux only.
--rseq-ops N
stop after N bogo rseq operations. Each bogo rseq operation is equivalent to 10000 iterations
over a long duration rseq handled critical section.
Real-time clock stressor
--rtc N
start N workers that exercise the real time clock (RTC) interfaces via /dev/rtc and
/sys/class/rtc/rtc0. No destructive writes (modifications) are performed on the RTC. This is
a Linux only stressor.
--rtc-ops N
stop after N bogo RTC interface accesses.
Fast process rescheduling stressor
--schedmix N
start N workers that each start child processes that repeatedly select random a scheduling
policy and then executes a short duration randomly chosen time consuming activity. This
exercises rapid re-scheduling of processes and generates a large amount of scheduling timer
interrupts.
--schedmix-ops N
stop after N scheduling mixed operations.
--schedmix-procs N
specify the number of chid processes to run for each stressor instance, range from 1 to 64,
default is 16.
Scheduling policy stressor
--schedpolicy N
start N workers that set the worker to various available scheduling policies out of
SCHED_OTHER, SCHED_BATCH, SCHED_IDLE, SCHED_FIFO, SCHED_RR, SCHED_DEADLINE and SCHED_EXT. For
the real time scheduling policies a random sched priority is selected between the minimum and
maximum scheduling priority settings.
--schedpolicy-ops N
stop after N bogo scheduling policy changes.
--schedpolicy-rand
Select scheduling policy randomly so that the new policy is always different to the previous
policy. The default is to work through the scheduling policies sequentially.
Stream control transmission protocol (SCTP) stressor
--sctp N
start N workers that perform network sctp stress activity using the Stream Control
Transmission Protocol (SCTP). This involves client/server processes performing rapid
connect, send/receives and disconnects on the local host.
--sctp-domain D
specify the domain to use, the default is ipv4. Currently ipv4 and ipv6 are supported.
--sctp-if NAME
use network interface NAME. If the interface NAME does not exist, is not up or does not
support the domain then the loopback (lo) interface is used as the default.
--sctp-ops N
stop sctp workers after N bogo operations.
--sctp-port P
start at sctp port P. For N sctp worker processes, ports P to (P × 4) − 1 are used for ipv4,
ipv6 domains and ports P to P − 1 are used for the unix domain.
--sctp-sched [ fc | fcfs | prio | rr | wfq ]
specify SCTP scheduler, one of fc (fair capacity), fcfs (first come first served, the
default), prio (priority), rr (round-robin) or wfq (weighted fair queueing)
File sealing (SEAL) stressor (Linux)
--seal N
start N workers that exercise the fcntl(2) SEAL commands on a small anonymous file created
using memfd_create(2). After each SEAL command is issued the stressor also sanity checks if
the seal operation has sealed the file correctly. (Linux only).
--seal-ops N
stop after N bogo seal operations.
Secure computing stressor
--seccomp N
start N workers that exercise Secure Computing system call filtering. Each worker creates
child processes that write a short message to /dev/null and then exits. 2% of the child
processes have a seccomp filter that disallows the write system call and hence it is killed
by seccomp with a SIGSYS. Note that this stressor can generate many audit log messages each
time the child is killed. Requires CAP_SYS_ADMIN to run.
--seccomp-ops N
stop seccomp stress workers after N seccomp filter tests.
Secret memory stressor (Linux ≥ 5.11)
--secretmem N
start N workers that mmap pages using file mapping off a memfd_secret file descriptor. Each
stress loop iteration will expand the mappable region by 3 pages using ftruncate and mmap and
touches the pages. The pages are then fragmented by unmapping the middle page and then
umapping the first and last pages. This tries to force page fragmentation and also trigger
out of memory (OOM) kills of the stressor when the secret memory is exhausted. Note this is
a Linux 5.11+ only stressor and the kernel needs to be booted with "secretmem=" option to
allocate a secret memory reservation.
--secretmem-ops N
stop secretmem stress workers after N stress loop iterations.
IO seek stressor
--seek N
start N workers that randomly seeks and performs 512 byte read/write I/O operations on a
file. The default file size is 16 GB.
--seek-ops N
stop seek stress workers after N bogo seek operations.
--seek-punch
punch randomly located 8 K holes into the file to cause more extents to force a more
demanding seek stressor, (Linux only).
--seek-size N
specify the size of the file in bytes. Small file sizes allow the I/O to occur in the cache,
causing greater CPU load. Large file sizes force more I/O operations to drive causing more
wait time and more I/O on the drive. One can specify the size in units of Bytes, KBytes,
MBytes and GBytes using the suffix b, k, m or g.
POSIX semaphore stressor
--sem N
start N workers that perform POSIX semaphore wait and post operations. By default, a parent
and 4 children are started per worker to provide some contention on the semaphore. This
stresses fast semaphore operations and produces rapid context switching.
--sem-ops N
stop semaphore stress workers after N bogo semaphore operations.
--sem-procs N
start N child workers per worker to provide contention on the semaphore, the default is 4 and
a maximum of 64 are allowed.
--sem-shared
share the semaphore across all sem stressor instances. Normally each semaphore stressor
shares a semaphore with its child processes, this option produces more locking contention and
less throughput by sharing a semaphore across all semaphore stressor processes.
System V semaphore stressor
--sem-sysv N
start N workers that perform System V semaphore wait and post operations. By default, a
parent and 4 children are started per worker to provide some contention on the semaphore.
This stresses fast semaphore operations and produces rapid context switching.
--sem-sysv-ops N
stop semaphore stress workers after N bogo System V semaphore operations.
--sem-sysv-procs N
start N child processes per worker to provide contention on the System V semaphore, the
default is 4 and a maximum of 64 are allowed.
--sem-sysv-setall
sets the semval values for all the semaphores in the child's semaphore set. This depends on
the semctl SETALL op being defined and GETALL succeeding and is an opt-in option as it will
affect the semaphore being exercised.
Sendfile stressor
--sendfile N
start N workers that send an empty file to /dev/null. This operation spends nearly all the
time in the kernel. The default sendfile size is 4 MB. The sendfile options are for Linux
only.
--sendfile-ops N
stop sendfile workers after N sendfile bogo operations.
--sendfile-size S
specify the size to be copied with each sendfile call. The default size is 4 MB. One can
specify the size in units of Bytes, KBytes, MBytes and GBytes using the suffix b, k, m or g.
Sessions stressor
--session N
start N workers that create child and grandchild processes that set and get their session
ids. 25% of the grandchild processes are not waited for by the child to create orphaned
sessions that need to be reaped by init.
--session-ops N
stop session workers after N child processes are spawned and reaped.
Setting data in the Kernel stressor
--set N
start N workers that call system calls that try to set data in the kernel, currently these
are: setgid, sethostname, setpgid, setpgrp, setuid, setgroups, setreuid, setregid, setresuid,
setresgid and setrlimit. Some of these system calls are OS specific.
--set-ops N
stop set workers after N bogo set operations.
Shellsort stressor
--shellsort N
start N workers that sort 32 bit integers using shellsort.
--shellsort-ops N
stop shellsort stress workers after N bogo shellsorts.
--shellsort-size N
specify number of 32 bit integers to sort, default is 262144 (256 × 1024).
POSIX shared memory stressor
--shm N
start N workers that open and allocate shared memory objects using the POSIX shared memory
interfaces. By default, the test will repeatedly create and destroy 32 shared memory
objects, each of which is 8 MB in size.
--shm-bytes N
specify the size of the POSIX shared memory objects to be created. One can specify the size
as % of total available memory or in units of Bytes, KBytes, MBytes and GBytes using the
suffix b, k, m or g.
--shm-mlock
attempt to mlock shared memory objects into memory causing more memory pressure by preventing
pages from swapped out.
--shm-objs N
specify the number of shared memory objects to be created.
--shm-ops N
stop after N POSIX shared memory create and destroy bogo operations are complete.
System V shared memory stressor
--shm-sysv N
start N workers that allocate shared memory using the System V shared memory interface. By
default, the test will repeatedly create and destroy 8 shared memory segments, each of which
is 8 MB in size.
--shm-sysv-bytes N
specify the size of the shared memory segment to be created. One can specify the size as % of
total available memory or in units of Bytes, KBytes, MBytes and GBytes using the suffix b, k,
m or g.
--shm-sysv-mlock
attempt to mlock shared memory segment into memory causing more memory pressure by preventing
pages from swapped out.
--shm-sysv-ops N
stop after N shared memory create and destroy bogo operations are complete.
--shm-sysv-segs N
specify the number of shared memory segments to be created. The default is 8 segments.
SIGABRT stressor
--sigabrt N
start N workers that create children that are killed by SIGABRT signals or by calling
abort(3).
--sigabrt-ops N
stop the sigabrt workers after N SIGABRT signals are successfully handled.
SIGBUS stressor
--sigbus N
start N workers that rapidly create and catch bus errors generated via misaligned access and
accessing a file backed memory mapping that does not have file storage to back the page being
accessed.
--sigbus-ops N
stop sigbus stress workers after N bogo bus errors.
SIGCHLD stressor
--sigchld N
start N workers that create children to generate SIGCHLD signals. This exercises children
that exit (CLD_EXITED), get killed (CLD_KILLED), get stopped (CLD_STOPPED) or continued
(CLD_CONTINUED).
--sigchld-ops N
stop the sigchld workers after N SIGCHLD signals are successfully handled.
SIGFD stressor (Linux)
--sigfd N
start N workers that generate SIGRT signals and are handled by reads by a child process using
a file descriptor set up using signalfd(2). (Linux only). This will generate a heavy context
switch load when all CPUs are fully loaded.
--sigfd-ops
stop sigfd workers after N bogo SIGUSR1 signals are sent.
SIGFPE stressor
--sigfpe N
start N workers that rapidly cause division by zero SIGFPE faults.
--sigfpe-ops N
stop sigfpe stress workers after N bogo SIGFPE faults.
SIGHUP stressor
--sighup N
start N workers that generate SIGHUP signals using raise(2) and by killing a group leader
process with a child process indirectly receiving SIGHUP when it loses the group leader.
--sighup-ops N
stop sighup stressor workers after N SIGHUP signals
SIGILL stressor
--sigill N
start N workers that execute illegal instructions to generate SIGILL signals.
--sigill-ops N
stop sigill stressor workers after N SIGILL signals
SIGIO stressor
--sigio N
start N workers that read data from a child process via a pipe and generate SIGIO signals.
This exercises asynchronous I/O via SIGIO.
--sigio-ops N
stop sigio stress workers after handling N SIGIO signals.
System signals stressor
--signal N
start N workers that exercise the signal system call three different signal handlers, SIG_IGN
(ignore), a SIGCHLD handler and SIG_DFL (default action). For the SIGCHLD handler, the
stressor sends itself a SIGCHLD signal and checks if it has been handled. For other handlers,
the stressor checks that the SIGCHLD handler has not been called. This stress test calls the
signal system call directly when possible and will try to avoid the C library attempt to
replace signal with the more modern sigaction system call.
--signal-ops N
stop signal stress workers after N rounds of signal handler setting.
Nested signal handling stressor
--signest N
start N workers that exercise nested signal handling. A signal is raised and inside the
signal handler a different signal is raised, working through a list of signals to exercise.
An alternative signal stack is used that is large enough to handle all the nested signal
calls. The -v option will log the approximate size of the stack required and the average
stack size per nested call.
--signest-ops N
stop after handling N nested signals.
Pending signals stressor
--sigpending N
start N workers that check if SIGUSR1 signals are pending. This stressor masks SIGUSR1,
generates a SIGUSR1 signal and uses sigpending(2) to see if the signal is pending. Then it
unmasks the signal and checks if the signal is no longer pending.
--sigpending-ops N
stop sigpending stress workers after N bogo sigpending pending/unpending checks.
SIGPIPE stressor
--sigpipe N
start N workers that repeatedly spawn off child process that exits before a parent can
complete a pipe write, causing a SIGPIPE signal. The child process is either spawned using
clone(2) if it is available or use the slower fork(2) instead.
--sigpipe-ops N
stop N workers after N SIGPIPE signals have been caught and handled.
Signal queueing stressor
--sigq N
start N workers that rapidly send SIGUSR1 signals using sigqueue(3) to child processes that
wait for the signal via sigwaitinfo(2).
--sigq-ops N
stop sigq stress workers after N bogo signal send operations.
Real-time signals stressor
--sigrt N
start N workers that each create child processes to handle SIGRTMIN to SIGRMAX real time
signals. The parent sends each child process a RT signal via siqueue(2) and the child process
waits for this via sigwaitinfo(2). When the child receives the signal it then sends a RT
signal to one of the other child processes also via sigqueue(2).
--sigrt-ops N
stop sigrt stress workers after N bogo sigqueue signal send operations.
SIGSEV stressor
--sigsegv N
start N workers that rapidly create and catch segmentation faults generated via illegal
memory access, illegal vdso system calls, illegal port reads, illegal interrupts or access to
x86 time stamp counter.
--sigsegv-ops N
stop sigsegv stress workers after N bogo segmentation faults.
Waiting for process signals stressor
--sigsuspend N
start N workers that each spawn off 4 child processes that wait for a SIGUSR1 signal from the
parent using sigsuspend(2). The parent sends SIGUSR1 signals to each child in rapid
succession. Each sigsuspend wakeup is counted as one bogo operation.
--sigsuspend-ops N
stop sigsuspend stress workers after N bogo sigsuspend wakeups.
SIGTRAP stressor
--sigtrap N
start N workers that exercise the SIGTRAP signal. For systems that support SIGTRAP, the
signal is generated using raise(SIGTRAP). Only x86 Linux systems the SIGTRAP is also
generated by an int 3 instruction.
--sigtrap-ops N
stop sigtrap stress workers after N SIGTRAP signals have been handled.
SIGURG stressor
--sigurg N
start workers that exercise the SIGURG signal by sending out-of-band data over a TCP/IP IPv4
socket stream.
--sigurg-ops N
stop sigurg stressor workers after N SIGURG signals have been handled.
SIGVTARLM stressor
--sigvtalrm N
start N workers that exercise the SIGVTALRM signal using an ITIMER_VIRTUAL itimer and a busy
loop that consumes CPU time calling getitimer for the ITIMER_VIRTUAL timer.
--sigvtalrm-ops N
stop sigvtalrm stress workers after N SIGVTALRM signals have been handled.
SIGXCPU stressor
--sigxcpu N
start N workers that exercise the SIGXCPU stressor. A busy loop generates SIGXCPU signals by
setting a 0 second run time limit followed by a sched_yield call.
--sigxcpu-ops N
stop sigsxcpu stress workers after N bogo SIGXCPU signal attempts.
SIGXFSZ stressor
--sigxfsz N
start N workers that exercise the SIGXFSZ stressor. A random 32 bit file size limit is set
and data is written outside this size limit to generate a SIGXFSZ signal.
--sigxfsz-ops N
stop sigsxfsz stress workers after N bogo SIGXFSZ signal attempts.
Random memory and processor cache line stressor via a skiplist
--skiplist N
start N workers that store and then search for integers using a skiplist. By default, 65536
integers are added and searched. This is a useful method to exercise random access of memory
and processor cache.
--skiplist-ops N
stop the skiplist worker after N skiplist store and search cycles are completed.
--skiplist-size N
specify the size (number of integers) to store and search in the skiplist. Size can be from 1
K to 4 M.
Time interrupts and context switches stressor
--sleep N
start N workers that spawn off multiple threads that each perform multiple sleeps of ranges
1us to 0.1s. This creates multiple context switches and timer interrupts.
--sleep-max P
start P threads per worker. The default is 1024, the maximum allowed is 30000.
--sleep-ops N
stop after N sleep bogo operations.
System management interrupts (SMI) stressor
--smi N
start N workers that attempt to generate system management interrupts (SMIs) into the x86
ring -2 system management mode (SMM) by exercising the advanced power management (APM) port
0xb2. This requires the --pathological option and root privilege and is only implemented on
x86 Linux platforms. This probably does not work in a virtualized environment. The stressor
will attempt to determine the time stolen by SMIs with some naïve benchmarking.
--smi-ops N
stop after N attempts to trigger the SMI.
Network socket stressor
-S N, --sock N
start N workers that perform various socket stress activity. This involves a pair of
client/server processes performing rapid connect, send and receives and disconnects on the
local host.
--sock-domain D
specify the domain to use, the default is ipv4. Currently ipv4, ipv6 and unix are supported.
--sock-if NAME
use network interface NAME. If the interface NAME does not exist, is not up or does not
support the domain then the loopback (lo) interface is used as the default.
--sock-msgs N
send N messages per connect, send/receive, disconnect iteration. The default is 1000
messages. If N is too small then the rate is throttled back by the overhead of socket connect
and disconnect (on Linux, one needs to increase /proc/sys/net/netfilter/nf_conntrack_max to
allow more connections).
--sock-nodelay
This disables the TCP Nagle algorithm, so data segments are always sent as soon as possible.
This stops data from being buffered before being transmitted, hence resulting in poorer
network utilisation and more context switches between the sender and receiver.
--sock-ops N
stop socket stress workers after N bogo operations.
--sock-opts [ random | send | sendmsg | sendmmsg ]
by default, messages are sent using send(2). This option allows one to specify the sending
method using send(2), sendmsg(2), sendmmsg(2) or a random selection of one of these 3 on each
iteration. Note that sendmmsg is only available for Linux systems that support this system
call.
--sock-port P
start at socket port P. For N socket worker processes, ports P to P − 1 are used.
--sock-protocol P
Use the specified protocol P, default is tcp. Options are tcp and mptcp (if supported by the
operating system).
--sock-type [ stream | seqpacket ]
specify the socket type to use. The default type is stream. seqpacket currently only works
for the unix socket domain.
--sock-zerocopy
enable zerocopy for send and recv calls if the MSG_ZEROCOPY is supported.
Socket abusing stressor
--sockabuse N
start N workers that abuse a socket file descriptor with various file based system that don't
normally act on sockets. The kernel should handle these illegal and unexpected calls
gracefully.
--sockabuse-ops N
stop after N iterations of the socket abusing stressor loop.
--sockabuse-port P
start at socket port P. For N sockabuse worker processes, ports P to P − 1 are used.
Socket diagnostic stressor (Linux)
--sockdiag N
start N workers that exercise the Linux sock_diag netlink socket diagnostics (Linux only).
This currently requests diagnostics using UDIAG_SHOW_NAME, UDIAG_SHOW_VFS, UDIAG_SHOW_PEER,
UDIAG_SHOW_ICONS, UDIAG_SHOW_RQLEN and UDIAG_SHOW_MEMINFO for the AF_UNIX family of socket
connections.
--sockdiag-ops N
stop after receiving N sock_diag diagnostic messages.
Socket file descriptor stressor
--sockfd N
start N workers that pass file descriptors over a UNIX domain socket using the CMSG(3)
ancillary data mechanism. For each worker, pair of client/server processes are created, the
server opens as many file descriptors on /dev/null as possible and passing these over the
socket to a client that reads these from the CMSG data and immediately closes the files.
--sockfd-ops N
stop sockfd stress workers after N bogo operations.
--sockfd-port P
start at socket port P. For N socket worker processes, ports P to P − 1 are used.
--sockfd-reuse
reuse the file descriptor by passing it back from the receiver to the sender and re-sending
it again rather than opening /dev/null each time.
Opening network socket stressor
--sockmany N
start N workers that use a client process to attempt to open as many as 100000 TCP/IP socket
connections to a server on port 10000.
--sockmany-if NAME
use network interface NAME. If the interface NAME does not exist, is not up or does not
support the domain then the loopback (lo) interface is used as the default.
--sockmany-ops N
stop after N connections.
--sockmany-port P
start at socket port P. For N sockmany worker processes, ports P to P − 1 are used.
Socket I/O stressor
--sockpair N
start N workers that perform socket pair I/O read/writes. This involves a pair of
client/server processes performing randomly sized socket I/O operations.
--sockpair-ops N
stop socket pair stress workers after N bogo operations.
Softlockup stressor
--softlockup N
start N workers that flip between with the "real-time" SCHED_FIO and SCHED_RR scheduling
policies at the highest priority to force softlockups. This can only be run with CAP_SYS_NICE
capability and for best results the number of stressors should be at least the number of
online CPUs. Once running, this is practically impossible to stop and it will force
softlockup issues and may trigger watchdog timeout reboots.
--softlockup-ops N
stop softlockup stress workers after N bogo scheduler policy changes.
Sparse matrix stressor
--sparsematrix N
start N workers that exercise 3 different sparse matrix implementations based on hashing,
Judy array (for 64 bit systems), 2-d circular linked-lists, memory mapped 2-d matrix (non-
sparse), quick hashing (on preallocated nodes) and red-black tree. The sparse matrix is
populated with values, random values potentially non-existing values are read, known existing
values are read and known existing values are marked as zero. This default 500 × 500 sparse
matrix is used and 5000 items are put into the sparse matrix making it 2% utilized.
--sparsematrix-items N
populate the sparse matrix with N items. If N is greater than the number of elements in the
sparse matrix than N will be capped to create at 100% full sparse matrix.
--sparsematrix-method [ all | hash | hashjudy | judy | list | mmap | qhash | rb ]
specify the type of sparse matrix implementation to use. The `all' method uses all the
methods and is the default.
Method Description
all exercise with all the sparsematrix stressor methods (see below):
hash use a hash table and allocate nodes on the heap for each unique value at a (x, y)
matrix position.
hashjudy use a hash table for x coordinates and a Judy array for y coordinates for values at
a (x, y) matrix position.
judy use a Judy array with a unique 1-to-1 mapping of (x, y) matrix position into the
array.
list use a circular linked-list for sparse y positions each with circular linked-lists
for sparse x positions for the (x, y) matrix coordinates.
mmap use a non-sparse mmap the entire 2-d matrix space. Only (x, y) matrix positions
that are referenced will get physically mapped. Note that large sparse matrices
cannot be mmap'd due to lack of virtual address limitations, and too many
referenced pages can trigger the out of memory killer on Linux.
qhash use a hash table with pre-allocated nodes for each unique value. This is a quick
hash table implementation, nodes are not allocated each time with calloc and are
allocated from a pre-allocated pool leading to quicker hash table performance than
the hash method.
rb use a red-black balanced tree using one tree node for each unique value at a (x, y)
matrix position.
splay use a splay tree using one tree node for each unique value at a (x, y) matrix
position.
--sparsematrix-ops N
stop after N sparsematrix test iterations.
--sparsematrix-size N
use a N × N sized sparse matrix
POSIX process spawn (posix_spawn) stressor (Linux)
--spawn N
start N workers continually spawn children using posix_spawn(3) that exec stress-ng and then
exit almost immediately. Currently Linux only.
--spawn-ops N
stop spawn stress workers after N bogo spawns.
Spinmem stressor
--spinmem N
start N workers that use a shared memory page to keep two processes synchronized using busy
spin loops. One process is a writer that increments a value in memory slot 0 and spin waits
for the data to appear in memory slot 1. The other process spin waits for the data to change
in slot 0 and copies the changed value into memory slot 1. The stressor benchmarks the time
for the transactions to occur between both processes. Note that the optimal number of
stressors is half the number of online CPUs in a system. By default 32 bit write/reads are
used.
--spinmem-affinity
move spinmem stressor processes to randomly selected CPUs every million spinmem transactions.
--spinmem-method [ 8bit | 16bit | 32bit | 64bit ]
select the size of the memory write/reads, the default is 32 bit.
--spinmem-numa
assign shared memory page to randomly selected NUMA nodes. This is disabled for systems that
do not support NUMA.
--spinmem-ops N
stop after N spinmem bogo operations. A bogo operation is 1000 transactions betewen the two
spinmem processes.
Splice stressor (Linux)
--splice N
move data from /dev/zero to /dev/null through a pipe without any copying between kernel
address space and user address space using splice(2). This is only available for Linux.
--splice-bytes N
transfer N bytes per splice call, the default is 64 K. One can specify the size as % of total
available memory or in units of Bytes, KBytes, MBytes and GBytes using the suffix b, k, m or
g.
--splice-ops N
stop after N bogo splice operations.
Stack stressor
--stack N
start N workers that rapidly cause and catch stack overflows by use of large recursive stack
allocations. Much like the brk stressor, this can eat up pages rapidly and may trigger the
kernel OOM killer on the process, however, the killed stressor is respawned again by a
monitoring parent process.
--stack-fill
the default action is to touch the lowest page on each stack allocation. This option touches
all the pages by filling the new stack allocation with zeros which forces physical pages to
be allocated and hence is more aggressive.
--stack-mlock
attempt to mlock stack pages into memory causing more memory pressure by preventing pages
from swapped out.
--stack-ops N
stop stack stress workers after N bogo stack overflows.
--stack-pageout
force stack pages out to swap (available when madvise(2) supports MADV_PAGEOUT).
--stack-unmap
unmap a single page in the middle of a large buffer allocated on the stack on each stack
allocation. This forces the stack mapping into multiple separate allocation mappings.
Dirty page and stack exception stressor
--stackmmap N
start N workers that use a 2 MB stack that is memory mapped onto a temporary file. A
recursive function works down the stack and flushes dirty stack pages back to the memory
mapped file using msync(2) until the end of the stack is reached (stack overflow). This
exercises dirty page and stack exception handling.
--stackmmap-ops N
stop workers after N stack overflows have occurred.
Statmount and listmount system call stressor
--statmount N
start N workers that find mounts on / via listmount and get mount information on the mount
IDs using statmount. (Linux only).
--statmount-ops N
stop workers after iterating on mounts N times.
Libc string functions stressor
--str N
start N workers that exercise various libc string functions on random strings.
--str-method strfunc
select a specific libc string function to stress. Available string functions to stress are:
all, index, rindex, strcasecmp, strcat, strchr, strcoll, strcmp, strcpy, strlen, strncasecmp,
strncat, strncmp, strrchr and strxfrm. See string(3) for more information on these string
functions. The `all' method is the default and will exercise all the string methods.
--str-ops N
stop after N bogo string operations.
STREAM memory stressor
--stream N
start N workers exercising a memory bandwidth stressor very loosely based on the STREAM
"Sustainable Memory Bandwidth in High Performance Computers" benchmarking tool by John D.
McCalpin, Ph.D. This stressor allocates buffers that are at least 4 times the size of the CPU
L2 cache and continually performs rounds of following computations on large arrays of double
precision floating point numbers:
Operation Description
copy c[i] = a[i]
scale b[i] = scalar × c[i]
add c[i] = a[i] + b[i]
triad a[i] = b[i] + (c[i] × scalar)
Since this is loosely based on a variant of the STREAM benchmark code, DO NOT submit results
based on this as it is intended to in stress-ng just to stress memory and compute and NOT
intended for STREAM accurate tuned or non-tuned benchmarking whatsoever. Use the official
STREAM benchmarking tool if you desire accurate and standardised STREAM benchmarks.
The stressor calculates the memory read rate, memory write rate and floating point operations
rate. These will differ from the maximum theoretical read/write/compute rates because of loop
overheads and the use of volatile pointers to ensure the compiler does not optimize out
stores.
--stream-index N
specify number of stream indices used to index into the data arrays a, b and c. This adds
indirection into the data lookup by using randomly shuffled indexing into the three data
arrays. Level 0 (no indexing) is the default, and 3 is where all 3 arrays are indexed via 3
different randomly shuffled indexes. The higher the index setting the more impact this has on
L1, L2 and L3 caching and hence forces higher memory read/write latencies.
--stream-l3-size N
Specify the CPU Level 3 cache size in bytes. One can specify the size in units of Bytes,
KBytes, MBytes and GBytes using the suffix b, k, m or g. If the L3 cache size is not
provided, then stress-ng will attempt to determine the cache size, and failing this, will
default the size to 4 MB.
--stream-mlock
attempt to mlock the stream buffers into memory to prevent them from being swapped out.
--stream-madvise [ collapse | hugepage | nohugepage | normal ]
Specify the madvise options used on the memory mapped buffer used in the stream stressor.
Non-linux systems will only have the `normal' madvise advice. The default is `normal'.
--stream-ops N
stop after N stream bogo operations, where a bogo operation is one round of copy, scale, add
and triad operations.
Swap partitions stressor (Linux)
--swap N
start N workers that add and remove small randomly sizes swap partitions (Linux only). Note
that if too many swap partitions are added then the stressors may exit with exit code 3 (not
enough resources). Requires CAP_SYS_ADMIN to run.
--swap-ops N
stop the swap workers after N swapon/swapoff iterations.
--swap-self
attempt to swap out pages of the stressor.
Context switching between mutually tied processes stressor
-s N, --switch N
start N workers that force context switching between two mutually blocking/unblocking tied
processes. By default message passing over a pipe is used, but different methods are
available.
--switch-freq F
run the context switching at the frequency of F context switches per second. Note that the
specified switch rate may not be achieved because of CPU speed and memory bandwidth
limitations.
--switch-method [ mq | pipe | sem-sysv ]
select the preferred context switch block/run synchronization method, these are as follows:
Method Description
mq use posix message queue with a 1 item size. Messages are passed between a sender
and receiver process.
pipe single character messages are passed down a single character sized pipe between a
sender and receiver process.
sem-sysv a SYSV semaphore is used to block/run two processes.
--switch-ops N
stop context switching workers after N bogo operations.
Symlink stressor
--symlink N
start N workers creating and removing symbolic links.
--symlink-ops N
stop symlink stress workers after N bogo operations.
--symlink-sync
sync dirty data and metadata to disk.
Partial file syncing (sync_file_range) stressor
--sync-file N
start N workers that perform a range of data syncs across a file using sync_file_range(2).
Three mixes of syncs are performed, from start to the end of the file, from end of the file
to the start, and a random mix. A random selection of valid sync types are used, covering the
SYNC_FILE_RANGE_WAIT_BEFORE, SYNC_FILE_RANGE_WRITE and SYNC_FILE_RANGE_WAIT_AFTER flag bits.
--sync-file-bytes N
specify the size of the file to be sync'd. One can specify the size as % of free space on the
file system in units of Bytes, KBytes, MBytes and GBytes using the suffix b, k, m or g.
--sync-file-ops N
stop sync-file workers after N bogo sync operations.
CPU synchronized loads stressor
--syncload N
start N workers that produce sporadic short lived loads synchronized across N stressor
processes. By default repeated cycles of 125ms busy load followed by 62.5ms sleep occur
across all the workers in step to create bursts of load to exercise C state transitions and
CPU frequency scaling. The busy load and sleeps have +/-10% jitter added to try exercising
scheduling patterns.
--syncload-msbusy M
specify the busy load duration in milliseconds.
--syncload-mssleep M
specify the sleep duration in milliseconds.
--syncload-ops N
stop syncload workers after N load/sleep cycles.
System calls bad address and fault handling stressor
--sysbadaddr N
start N workers that pass bad addresses to system calls to exercise bad address and fault
handling. The addresses used are null pointers, read only pages, write only pages, unmapped
addresses, text only pages, unaligned addresses and top of memory addresses.
--sysbadaddr-ops N
stop the sysbadaddr stressors after N bogo system calls.
System calls stressor
--syscall N
start N workers that exercise a range of available system calls. System calls that fail due
to lack of capabilities or errors are ignored. The stressor will try to maximize the rate of
system calls being executed based the entire time taken to setup, run and cleanup after each
system call.
--syscall-method method
select the choice of system calls to executed based on the fastest test duration times. Note
that this includes the time to setup, execute the system call and cleanup afterwards. The
available methods are as follows:
Method Description
all select all the available system calls
fast10 select the fastest 10% system call tests
fast25 select the fastest 25% system call tests
fast50 select the fastest 50% system call tests
fast75 select the fastest 75% system call tests
fast90 select the fastest 90% system call tests
geomean1 select tests that are less or equal to the geometric mean of all the test times
geomean1 select tests that are less or equal to 2 × the geometric mean of all the test
times
geomean1 select tests that are less or equal to 3 × the geometric mean of all the test
times
--syscall-ops N
stop after N system calls
--sycsall-top N
report the fastest top N system calls. Setting N to zero will report all the system calls
that could be exercised.
System information stressor
--sysinfo N
start N workers that continually read system and process specific information. This reads
the process user and system times using the times(2) system call. For Linux systems, it also
reads overall system statistics using the sysinfo(2) system call and also the file system
statistics for all mounted file systems using statfs(2).
--sysinfo-ops N
stop the sysinfo workers after N bogo operations.
System calls with invalid arguments stressor (Linux)
--sysinval N
start N workers that exercise system calls in random order with permutations of invalid
arguments to force kernel error handling checks. The stress test autodetects system calls
that cause processes to crash or exit prematurely and will blocklist these after several
repeated breakages. System call arguments that cause system calls to work successfully are
also detected an blocklisted too. Linux only.
--sysinval-ops N
stop sysinval workers after N system call attempts.
/sys stressor (Linux)
--sysfs N
start N workers that recursively read files from /sys (Linux only). This may cause specific
kernel drivers to emit messages into the kernel log.
--sysfs-ops N
stop sysfs reading after N bogo read operations. Note, since the number of entries may vary
between kernels, this bogo ops metric is probably very misleading.
Tee stressor (Linux)
--tee N
move data from a writer process to a reader process through pipes and to /dev/null without
any copying between kernel address space and user address space using tee(2). This is only
available for Linux.
--tee-ops N
stop after N bogo tee operations.
Timer event stressor (Linux)
-T N, --timer N
start N workers creating timer events at a default rate of 1 MHz (Linux only); this can
create a many thousands of timer clock interrupts. Each timer event is caught by a signal
handler and counted as a bogo timer op.
--timer-freq F
run timers at F Hz; range from 1 to 1000000000 Hz (Linux only). By selecting an appropriate
frequency stress-ng can generate hundreds of thousands of interrupts per second. Note: it is
also worth using --timer-slack 0 for high frequencies to stop the kernel from coalescing
timer events.
--timer-ops N
stop timer stress workers after N bogo timer events (Linux only).
--timer-rand
select a timer frequency based around the timer frequency +/- 12.5% random jitter. This tries
to force more variability in the timer interval to make the scheduling less predictable.
Timerfd stressor (Linux)
--timerfd N
start N workers creating timerfd events at a default rate of 1 MHz (Linux only); this can
create a many thousands of timer clock events. Timer events are waited for on the timer file
descriptor using select(2) and then read and counted as a bogo timerfd op.
--timerfs-fds N
try to use a maximum of N timerfd file descriptors per stressor.
--timerfd-freq F
run timers at F Hz; range from 1 to 1000000000 Hz (Linux only). By selecting an appropriate
frequency stress-ng can generate hundreds of thousands of interrupts per second.
--timerfd-ops N
stop timerfd stress workers after N bogo timerfd events (Linux only).
--timerfd-rand
select a timerfd frequency based around the timer frequency +/- 12.5% random jitter. This
tries to force more variability in the timer interval to make the scheduling less
predictable.
Time warp stressor
--time-warp N
start N workers that read the system time and where appropriate for monotonic clocks perform
a check for reverse time warping. At the end of the run there are time wrap-around checks.
This stressor exercises clock_gettime(2) on all available clocks, gettimeofday(2), time(2)
and getrusage(2) to check for unexpected time behaviour. Note that only some clocks are
reliably monotonic.
--time-warp-ops N
stop after N rounds of checking all the available clocks and time fetching system calls.
Translation lookaside buffer shootdowns stressor
--tlb-shootdown N
start N workers that force Translation Lookaside Buffer (TLB) shootdowns. This is achieved
by creating up to 16 child processes that all share a region of memory and these processes
are shared amongst the available CPUs. The processes adjust the page mapping settings
causing TLBs to be force flushed on the other processors, causing the TLB shootdowns.
--tlb-shootdown-ops N
stop after N bogo TLB shootdown operations are completed.
Tmpfs stressor
--tmpfs N
start N workers that create a temporary file on an available tmpfs file system and perform
various file based mmap operations upon it.
--tmpfs-mmap-async
enable file based memory mapping and use asynchronous msync'ing on each page, see
--tmpfs-mmap-file.
--tmpfs-mmap-file
enable tmpfs file based memory mapping and by default use synchronous msync'ing on each page.
--tmpfs-ops N
stop tmpfs stressors after N bogo mmap operations.
Touching files stressor
--touch N
touch files by using open(2) or creat(2) and then closing and unlinking them. The filename
contains the bogo-op number and is incremented on each touch operation, hence this fills the
dentry cache. Note that the user time and system time may be very low as most of the run time
is waiting for file I/O and this produces very large bogo-op rates for the very low CPU time
used.
--touch-method [ random | open | creat ]
select the method the file is created, either randomly using open(2) or create(2), just using
open(2) with the O_CREAT open flag, or with creat(2).
--touch-ops N
stop the touch workers after N file touches.
--touch-opts all, direct, dsync, excl, noatime, sync
specify various file open options as a comma separated list. Options are as follows:
Option Description
all use all the open options, namely direct, dsync, excl, noatime and sync
direct try to minimize cache effects of the I/O to and from this file, using the O_DIRECT
open flag.
dsync ensure output has been transferred to underlying hardware and file metadata has
been updated using the O_DSYNC open flag.
excl fail if file already exists (it should not).
noatime do not update the file last access time if the file is read.
sync ensure output has been transferred to underlying hardware using the O_SYNC open
flag.
Tree data structures stressor
--tree N
start N workers that exercise tree data structures. The default is to add, find and remove
250,000 64 bit integers into AVL (avl), Red-Black (rb), Splay (splay), btree and binary
trees. The intention of this stressor is to exercise memory and cache with the various tree
operations.
--tree-method [ all | avl | binary | btree | rb | splay ]
specify the tree to be used. By default, all the trees are used (the `all' option).
--tree-ops N
stop tree stressors after N bogo ops. A bogo op covers the addition, finding and removing all
the items into the tree(s).
--tree-size N
specify the size of the tree, where N is the number of 64 bit integers to be added into the
tree.
Trigonometric functions stressor
--trig N
start N workers that exercise sin, cos, sincos (where available) and tan trigonometric
functions using float, double and long double floating point variants. Each function is
exercised 10,000 times per bogo-operation.
--trig-method function
specify a trigonometric stress function. By default, all the functions are exercised
sequentially, however one can specify just one function to be used if required. Available
options are as follows:
Method Description
all iterate through all of the following trigonometric functions
cos cosine (double precision)
cosf cosine (float precision)
cosl cosine (long double precision)
sin sine (double precision)
sinf sine (float precision)
sinl sine (long double precision)
sincos sine and cosine (double precision)
sincosf sine and cosine (float precision)
sincosl sine and cosine (long double precision)
tan tangent (double precision)
tanf tangent (float precision)
tanl tangent (long double precision)
--trig-ops N
stop after N bogo-operations.
Time stamp counter (TSC) stressor
--tsc N
start N workers that read the Time Stamp Counter (TSC) 256 times per loop iteration (bogo
operation). This exercises the tsc instruction for x86, the mftb instruction for ppc64, the
rdcycle instruction for RISC-V, the tick instruction on SPARC and the rdtime.d instruction
for Loong64.
--tsc-lfence
add lfence after each tsc read to force serialization (x86 only).
--tsc-ops N
stop the tsc workers after N bogo operations are completed.
--tsc-rdtscp
use the rdtscp instruction instead of rdtsc (x86 only). This also disables the --tsc-lfence
option.
Binary tree stressor
--tsearch N
start N workers that insert, search and delete 32 bit integers on a binary tree using
tsearch(3), tfind(3) and tdelete(3). By default, there are 65536 randomized integers used in
the tree. This is a useful method to exercise random access of memory and processor cache.
--tsearch-ops N
stop the tsearch workers after N bogo tree operations are completed.
--tsearch-size N
specify the size (number of 32 bit integers) in the array to tsearch. Size can be from 1 K to
4 M.
Network tunnel stressor
--tun N
start N workers that create a network tunnel device and sends and receives packets over the
tunnel using UDP and then destroys it. A new random 192.168.*.* IPv4 address is used each
time a tunnel is created.
--tun-ops N
stop after N iterations of creating/sending/receiving/destroying a tunnel.
--tun-tap
use network tap device using level 2 frames (bridging) rather than a tun device for level 3
raw packets (tunnelling).
UDP network stressor
--udp N
start N workers that transmit data using UDP. This involves a pair of client/server processes
performing rapid connect, send and receives and disconnects on the local host.
--udp-domain D
specify the domain to use, the default is ipv4. Currently ipv4 and ipv6 are supported.
--udp-gro
enable UDP-GRO (Generic Receive Offload) if supported.
--udp-if NAME
use network interface NAME. If the interface NAME does not exist, is not up or does not
support the domain then the loopback (lo) interface is used as the default.
--udp-lite
use the UDP-Lite (RFC 3828) protocol (only for ipv4 and ipv6 domains).
--udp-ops N
stop udp stress workers after N bogo operations.
--udp-port P
start at port P. For N udp worker processes, ports P to P − 1 are used. By default, ports
7000 upwards are used.
UDP flooding stressor
--udp-flood N
start N workers that attempt to flood the host with UDP packets to random ports. The IP
address of the packets are currently not spoofed. This is only available on systems that
support AF_PACKET.
--udp-flood-domain D
specify the domain to use, the default is ipv4. Currently ipv4 and ipv6 are supported.
--udp-flood-if NAME
use network interface NAME. If the interface NAME does not exist, is not up or does not
support the domain then the loopback (lo) interface is used as the default.
--udp-flood-ops N
stop udp-flood stress workers after N bogo operations.
File umask stressor
--umask N
start N workers that exercise setting umask and creating/fstat'ing/closing/unlinking a file
and checking the file mode is set correctly. This exercises umask mask values from octal 0000
to 0777 inclusive.
--umask-ops N
stop after N rounds of exercising umask values 0000 to 0777.
Umount stressor
--umount N
start N workers that exercise mounting and racying unmounting of small tmpfs and ramfs file
systems. Three child processes are invoked, one to mount, another to force umount and a third
to exercise /proc/mounts. Small random delays are used between mount and umount calls to try
to trigger race conditions on the umount calls.
--umount-ops N
stop umount workers after N successful bogo mount/umount operations.
Unlink stressor
--unlink N
start N workers that each run 4 processes per worker that create, open, unlink and close 1024
files in randomized order to exercise unlinking (removal) of files. This attempts to create
races on file creation, linking and unlinking.
--unlink-ops N
stop after N bogo rounds of unlink operations on 1024 files by the controlling worker.
Unshare stressor (Linux)
--unshare N
start N workers that each fork off 32 child processes, each of which exercises the unshare(2)
system call by disassociating parts of the process execution context. (Linux only).
--unshare-ops N
stop after N bogo unshare operations.
Uprobe stressor (Linux)
--uprobe N
start N workers that trace the entry to libc function getpid() using the Linux uprobe kernel
tracing mechanism. This requires CAP_SYS_ADMIN capabilities and a modern Linux uprobe capable
kernel.
--uprobe-ops N
stop uprobe tracing after N trace events of the function that is being traced.
/dev/urandom stressor (Linux)
-u N, --urandom N
start N workers reading /dev/urandom (Linux only). This will load the kernel random number
source.
--urandom-ops N
stop urandom stress workers after N urandom bogo read operations (Linux only).
Page faults stressor (Linux)
--userfaultfd N
start N workers that generate write page faults on a small anonymously mapped memory region
and handle these faults using the user space fault handling via the userfaultfd mechanism.
This will generate a large quantity of major page faults and also context switches during the
handling of the page faults. (Linux only).
--userfaultfd-bytes N
mmap N bytes per userfaultfd worker to page fault on, the default is 16 MB. One can specify
the size as % of total available memory or in units of Bytes, KBytes, MBytes and GBytes using
the suffix b, k, m or g.
--userfaultfd-ops N
stop userfaultfd stress workers after N page faults.
SYGSYS stressor
--usersyscall N
start N workers that exercise the Linux prctl userspace system call mechanism. A userspace
system call is handled by a SIGSYS signal handler and exercised with the system call disabled
(ENOSYS) and enabled (via SIGSYS) using prctl PR_SET_SYSCALL_USER_DISPATCH.
--usersyscall-ops N
stop after N successful userspace syscalls via a SIGSYS signal handler.
File timestamp stressor
--utime N
start N workers updating file timestamps. This is mainly CPU bound when the default is used
as the system flushes metadata changes only periodically.
--utime-fsync
force metadata changes on each file timestamp update to be flushed to disk. This forces the
test to become I/O bound and will result in many dirty metadata writes.
--utime-ops N
stop utime stress workers after N utime bogo operations.
Virtual dynamic shared object stressor
--vdso N
start N workers that repeatedly call each of the system call functions in the vDSO (virtual
dynamic shared object). The vDSO is a shared library that the kernel maps into the address
space of all user-space applications to allow fast access to kernel data to some system calls
without the need of performing an expensive system call.
--vdso-func F
Instead of calling all the vDSO functions, just call the vDSO function F. The functions
depend on the kernel being used, but are typically clock_gettime, getcpu, gettimeofday and
time.
--vdso-ops N
stop after N vDSO functions calls.
Vector integer comparison operations stressor
--veccmp N
start N workers that perform various unsigned integer math comparison operations on various
128 bit vectors. A mix of integer vector comparison operations are performed on the following
vectors: 16 × 8 bits, 8 × 16 bits, 4 × 32 bits, 2 × 64 bits and if supported 1 × 128 bits.
The metrics produced by this mix depend on the processor architecture and the vector math
optimisations produced by the compiler.
--veccmp-ops N
stop after N bogo vector integer comparison operations.
Vector floating point operations stressor
--vecfp N
start N workers that exericise floating point (single and double precision) addition,
multiplication, division and negation on vectors of 128, 64, 32, 16 and 8 floating point
values. The -v option will show the approximate throughput in millions of floating pointer
operations per second for each operation. For x86, the gcc/clang target clones attribute has
been used to produced vector optimizations for a range of mmx, sse, avx and processor
features.
--vecfp-method method
specify a vecfp stress method. By default, all the stress methods are exercised sequentially,
however one can specify just one method to be used if required.
Method Description
all iterate through all of the following vector methods
floatv128add addition of a vector of 128 single precision floating point values
floatv64add addition of a vector of 64 single precision floating point values
floatv32add addition of a vector of 32 single precision floating point values
floatv16add addition of a vector of 16 single precision floating point values
floatv8add addition of a vector of 8 single precision floating point values
floatv128mul multiplication of a vector of 128 single precision floating point values
floatv64mul multiplication of a vector of 64 single precision floating point values
floatv32mul multiplication of a vector of 32 single precision floating point values
floatv16mul multiplication of a vector of 16 single precision floating point values
floatv8mul multiplication of a vector of 8 single precision floating point values
floatv128div division of a vector of 128 single precision floating point values
floatv64div division of a vector of 64 single precision floating point values
floatv32div division of a vector of 32 single precision floating point values
floatv16div division of a vector of 16 single precision floating point values
floatv8div division of a vector of 8 single precision floating point values
doublev128add addition of a vector of 128 double precision floating point values
doublev64add addition of a vector of 64 double precision floating point values
doublev32add addition of a vector of 32 double precision floating point values
doublev16add addition of a vector of 16 double precision floating point values
doublev8add addition of a vector of 8 double precision floating point values
doublev128mul multiplication of a vector of 128 double precision floating point values
doublev64mul multiplication of a vector of 64 double precision floating point values
doublev32mul multiplication of a vector of 32 double precision floating point values
doublev16mul multiplication of a vector of 16 double precision floating point values
doublev8mul multiplication of a vector of 8 double precision floating point values
doublev128div division of a vector of 128 double precision floating point values
doublev64div division of a vector of 64 double precision floating point values
doublev32div division of a vector of 32 double precision floating point values
doublev16div division of a vector of 16 double precision floating point values
doublev8div division of a vector of 8 double precision floating point values
doublev128neg negation of a vector of 128 double precision floating point values
doublev64neg negation of a vector of 64 double precision floating point values
doublev32neg negation of a vector of 32 double precision floating point values
doublev16neg negation of a vector of 16 double precision floating point values
doublev8neg negation of a vector of 8 double precision floating point values
--vecfp-ops N
stop after N vector floating point bogo-operations. Each bogo-op is equivalent to 65536 loops
of 2 vector operations. For example, one bogo-op on a 16 wide vector is equivalent to 65536 ×
2 × 16 floating point operations.
Vector math operations stressor
--vecmath N
start N workers that perform various unsigned integer math operations on various 128 bit
vectors. A mix of vector math operations are performed on the following vectors: 16 × 8 bits,
8 × 16 bits, 4 × 32 bits, 2 × 64 bits and where supported 1 × 128 bits. The metrics produced
by this mix depend on the processor architecture and the vector math optimisations produced
by the compiler.
--vecmath-ops N
stop after N bogo vector integer math operations.
Shuffled vector math operations stressor
--vecshuf N
start N workers that shuffle data on various 64 byte vectors comprised of 8, 16, 32, 64 and
128 bit unsigned integers. The integers are shuffled around the vector with 4 shuffle
operations per loop, 65536 loops make up one bogo-op of shuffling. The data shuffling rates
and shuffle operation rates are logged when using the -v option. This stressor exercises
vector load, shuffle/permute, packing/unpacking and store operations.
--vecshuf-method method
specify a vector shuffling stress method. By default, all the stress methods are exercised
sequentially, however one can specify just one method to be used if required.
Method Description
all iterate through all of the following vector methods
u8x64 shuffle a vector of 64 unsigned 8 bit integers
u16x32 shuffle a vector of 32 unsigned 16 bit integers
u32x16 shuffle a vector of 16 unsigned 32 bit integers
u64x8 shuffle a vector of 8 unsigned 64 bit integers
u128x4 shuffle a vector of 4 unsigned 128 bit integers (when supported)
--vecshuf-ops N
stop after N bogo vector shuffle ops. One bogo-op is equavlent of 4 × 65536 vector shuffle
operations on 64 bytes of vector data.
Wide vector math operations stressor
--vecwide N
start N workers that perform various 8 bit math operations on vectors of 4, 8, 16, 32, 64,
128, 256, 512, 1024 and 2048 bytes. With the -v option the relative compute performance vs
the expected compute performance based on total run time is shown for the first vecwide
worker. The vecwide stressor exercises various processor vector instruction mixes and how
well the compiler can map the vector operations to the target instruction set.
--vecwide-ops N
stop after N bogo vector operations (2048 iterations of a mix of vector instruction
operations).
File based authenticy protection (verity) stressor
--verity N
start N workers that exercise read-only file based authenticy protection using the verity
ioctls FS_IOC_ENABLE_VERITY and FS_IOC_MEASURE_VERITY. This requires file systems with
verity support (currently ext4 and f2fs on Linux) with the verity feature enabled. The test
attempts to creates a small file with multiple small extents and enables verity on the file
and verifies it. It also checks to see if the file has verity enabled with the FS_VERITY_FL
bit set on the file flags.
--verity-ops N
stop the verity workers after N file create, enable verity, check verity and unlink cycles.
vfork stressor
--vfork N
start N workers continually vforking children that immediately exit.
--vfork-max P
create P processes and then wait for them to exit per iteration. The default is just 1;
higher values will create many temporary zombie processes that are waiting to be reaped. One
can potentially fill up the process table using high values for --vfork-max and --vfork.
--vfork-ops N
stop vfork stress workers after N bogo operations.
vfork processes as much as possible stressor
--vforkmany N
start N workers that spawn off a chain of vfork children until the process table fills up
and/or vfork fails. vfork can rapidly create child processes and the parent process has to
wait until the child dies, so this stressor rapidly fills up the process table.
--vforkmany-ops N
stop vforkmany stressors after N vforks have been made.
--vforkmany-vm
enable detrimental performance virtual memory advice using madvise on all pages of the
vforked process. Where possible this will try to set every page in the new process with using
madvise MADV_MERGEABLE, MADV_WILLNEED, MADV_HUGEPAGE and MADV_RANDOM flags. Linux only.
--vforkmany-vm-bytes N
mmap N bytes per vm worker for more memory pressure, the default is 64 MB. This also enables
the --vforkmany-vm option. One can specify the size as % of total available memory or in
units of Bytes, KBytes, MBytes and GBytes using the suffix b, k, m or g.
Memory allocate and write stressor
-m N, --vm N
start N workers continuously calling mmap(2)/munmap(2) and writing to the allocated memory.
Note that this can cause systems to trip the kernel OOM killer on Linux systems if not enough
physical memory and swap is not available.
--vm-bytes N
mmap N bytes in total, this is shared by each vm worker, the default is 256 MB. One can
specify the size as % of total available memory or in units of Bytes, KBytes, MBytes and
GBytes using the suffix b, k, m or g.
--vm-flush
cache flush mapped memory after each memory region has been completely written to.
--vm-hang N
sleep N seconds before unmapping memory, the default is zero seconds. Specifying 0 will do
an infinite wait.
--vm-keep
do not continually unmap and map memory, just keep on re-writing to it.
--vm-locked
Lock the pages of the mapped region into memory using mmap MAP_LOCKED (since Linux 2.5.37).
This is similar to locking memory as described in mlock(2).
--vm-madvise advice
Specify the madvise `advice' option used on the memory mapped regions used in the vm
stressor. Non-linux systems will only have the `normal' madvise advice, linux systems support
`collapse', `dontneed', `hugepage', `mergeable' , `nohugepage', `normal', `random',
`sequential', `unmergeable' and `willneed' advice. If this option is not used then the
default is to pick random madvise advice for each mmap call. See madvise(2) for more details.
--vm-method method
specify a vm stress method. By default, all the stress methods are exercised sequentially,
however one can specify just one method to be used if required. Each of the vm workers have
3 phases:
1. Initialised. The anonymously memory mapped region is set to a known pattern.
2. Exercised. Memory is modified in a known predictable way. Some vm workers alter memory
sequentially, some use small or large strides to step along memory.
3. Checked. The modified memory is checked to see if it matches the expected result.
The vm methods containing `prime' in their name have a stride of the largest prime less than
2↑64, allowing to them to thoroughly step through memory and touch all locations just once
while also doing without touching memory cells next to each other. This strategy exercises
the cache and page non-locality.
Since the memory being exercised is virtually mapped then there is no guarantee of touching
page addresses in any particular physical order. These workers should not be used to test
that all the system's memory is working correctly either, use tools such as memtest86
instead.
The vm stress methods are intended to exercise memory in ways to possibly find memory issues
and to try to force thermal errors.
Available vm stress methods are described as follows:
Method Description
all iterate over all the vm stress methods as listed below.
cache-lines work through memory in 64 byte cache sized steps writing a single byte per
cache line. Once the write is complete, the memory is read to verify the values
are written correctly.
cache-stripe work through memory in 64 byte cache sized chunks, writing in ascending address
order on even offsets and descending address order on odd offsets.
checkboard work through memory writing alternative zero/one bit values into memory in a
mixed checkerboard pattern. Memory is swapped around to ensure every bit is
read, bit flipped and re-written and then re-read for verification.
flip sequentially work through memory 8 times, each time just one bit in memory
flipped (inverted). This will effectively invert each byte in 8 passes.
fwdrev write to even addressed bytes in a forward direction and odd addressed bytes in
reverse direction. rhe contents are sanity checked once all the addresses have
been written to.
galpat-0 galloping pattern zeros. This sets all bits to 0 and flips just 1 in 4096 bits
to 1. It then checks to see if the 1s are pulled down to 0 by their neighbours
or of the neighbours have been pulled up to 1.
galpat-1 galloping pattern ones. This sets all bits to 1 and flips just 1 in 4096 bits
to 0. It then checks to see if the 0s are pulled up to 1 by their neighbours or
of the neighbours have been pulled down to 0.
gray fill the memory with sequential gray codes (these only change 1 bit at a time
between adjacent bytes) and then check if they are set correctly.
grayflip fill memory with adjacent bytes of gray code and inverted gray code pairs to
change as many bits at a time between adjacent bytes and check if these are set
correctly.
incdec work sequentially through memory twice, the first pass increments each byte by
a specific value and the second pass decrements each byte back to the original
start value. The increment/decrement value changes on each invocation of the
stressor.
inc-nybble initialise memory to a set value (that changes on each invocation of the
stressor) and then sequentially work through each byte incrementing the bottom
4 bits by 1 and the top 4 bits by 15.
lfsr32 fill memory with values generated from a 32 bit Galois linear feedback shift
register using the polynomial x↑32 + x↑31 + x↑29 + x + 1. This generates a
ring of 2↑32 − 1 unique values (all 32 bit values except for 0).
modulo-x fill memory over 23 iterations. Each iteration starts one byte further along
from the start of the memory and steps along in 23 byte strides. In each
stride, the first byte is set to a random pattern and all other bytes are set
to the inverse. Then it checks see if the first byte contains the expected
random pattern. This exercises cache store/reads as well as seeing if
neighbouring cells influence each other.
move-inv sequentially fill memory 64 bits of memory at a time with random values, and
then check if the memory is set correctly. Next, sequentially invert each 64
bit pattern and again check if the memory is set as expected.
mscan fill each bit in each byte with 1s then check these are set, fill each bit in
each byte with 0s and check these are clear.
one-zero set all memory bits to one and then check if any bits are not one. Next, set
all the memory bits to zero and check if any bits are not zero.
prime-0 iterate 8 times by stepping through memory in very large prime strides clearing
just on bit at a time in every byte. Then check to see if all bits are set to
zero.
prime-1 iterate 8 times by stepping through memory in very large prime strides setting
just on bit at a time in every byte. Then check to see if all bits are set to
one.
prime-gray-0 first step through memory in very large prime strides clearing just on bit
(based on a gray code) in every byte. Next, repeat this but clear the other 7
bits. Then check to see if all bits are set to zero.
prime-gray-1 first step through memory in very large prime strides setting just on bit
(based on a gray code) in every byte. Next, repeat this but set the other 7
bits. Then check to see if all bits are set to one.
prime-incdec step through memory in large prime steps incrementing bytes and then re-do
again decrementing bytes.
rand-set sequentially work through memory in 64 bit chunks setting bytes in the chunk to
the same 8 bit random value. The random value changes on each chunk. Check
that the values have not changed.
rand-sum sequentially set all memory to random values and then summate the number of
bits that have changed from the original set values.
read64 sequentially read memory using 32 × 64 bit reads per bogo loop. Each loop
equates to one bogo operation. This exercises raw memory reads.
ror fill memory with a random pattern and then sequentially rotate 64 bits of
memory right by one bit, then check the final load/rotate/stored values.
rowhammer try to force memory corruption using the rowhammer memory stressor. This
fetches two 32 bit integers from memory and forces a cache flush on the two
addresses multiple times. This has been known to force bit flipping on some
hardware, especially with lower frequency memory refresh cycles.
swap fill memory in 64 byte chunks with random patterns. Then swap each 64 chunk
with a randomly chosen chunk. Finally, reverse the swap to put the chunks back
to their original place and check if the data is correct. This exercises
adjacent and random memory load/stores.
walk-0a in the given memory mapping, work through a range of specially chosen addresses
working through address lines to see if any address lines are stuck low. This
works best with physical memory addressing, however, exercising these virtual
addresses has some value too.
walk-0d for each byte in memory, walk through each data line setting them to low (and
the others are set high) and check that the written value is as expected. This
checks if any data lines are stuck.
walk-1a in the given memory mapping, work through a range of specially chosen addresses
working through address lines to see if any address lines are stuck high. This
works best with physical memory addressing, however, exercising these virtual
addresses has some value too.
walk-1d for each byte in memory, walk through each data line setting them to high (and
the others are set low) and check that the written value is as expected. This
checks if any data lines are stuck.
walk-flush walk through memory a byte at a time writing/flushing/reading 8 bytes of
incrementing 8 bit data. For architectures that support user space cache
flushing this will cause high data cache miss rates.
write64 sequentially write to memory using 32 × 64 bit writes per bogo loop. Each loop
equates to one bogo operation. This exercises raw memory writes. Note that
memory writes are not checked at the end of each test iteration.
write64ds sequentially write to memory using 32 × 64 bit direct store writes per bogo
loop. Each loop equates to one bogo operation. This exercises cacheless write
combining (WC) memory type protocol for memory writes and is only available on
x86 systems with the movediri instruction built with gcc and clang compilers.
Note that memory writes are not checked at the end of each test iteration.
write64nt sequentially write to memory using 32 × 64 bit non-temporal writes per bogo
loop. Each loop equates to one bogo operation. This exercises cacheless raw
memory writes and is only available on x86 sse2 capable systems built with gcc
and clang compilers. Note that memory writes are not checked at the end of each
test iteration.
write1024v sequentially write to memory using 1 × 1024 bit vector write per bogo loop
(only available if the compiler supports vector types). Each loop equates to
one bogo operation. This exercises raw memory writes. Note that memory writes
are not checked at the end of each test iteration.
wrrd128nt write to memory in 128 bit chunks using non-temporal writes (bypassing the
cache). Each chunk is written 4 times to hammer the memory. Then check to see
if the data is correct using non-temporal reads if they are available or normal
memory reads if not. Only available with processors that provide non-temporal
128 bit writes.
zero-one set all memory bits to zero and then check if any bits are not zero. Next, set
all the memory bits to one and check if any bits are not one.
--vm-numa
assign memory mapped pages to randomly selected NUMA nodes. This is disabled for systems that
do not support NUMA.
--vm-ops N
stop vm workers after N bogo operations.
--vm-populate
populate (prefault) page tables for the memory mappings; this can stress swapping. Only
available on systems that support MAP_POPULATE (since Linux 2.5.46).
Virtual memory addressing stressor
--vm-addr N
start N workers that exercise virtual memory addressing using various methods to walk through
a memory mapped address range. This will exercise mapped private addresses from 8 MB to 64 MB
per worker and try to generate cache and TLB inefficient addressing patterns. Each method
will set the memory to a random pattern in a write phase and then sanity check this in a read
phase.
--vm-addr-method method
specify a vm address stress method. By default, all the stress methods are exercised
sequentially, however one can specify just one method to be used if required.
Available vm address stress methods are described as follows:
Method Description
all iterate over all the vm stress methods as listed below.
bitposn iteratively write to memory in powers of 2 strides of max_stride to 1 and then read
check memory in powers of 2 strides 1 to max_stride where max_stride is half the
size of the memory mapped region. All bit positions of the memory address space are
bit flipped in the striding.
dec work through the address range backwards sequentially, byte by byte.
decinv like dec, but with all the relevant address bits inverted.
flip address memory using gray coded addresses and their inverse to flip as many address
bits per write/read operation
gray work through memory with gray coded addresses so that each change of address just
changes 1 bit compared to the previous address.
grayinv like gray, but with the all relevant address bits inverted, hence all bits change
apart from 1 in the address range.
inc work through the address range forwards sequentially, byte by byte.
incinv like inc, but with all the relevant address bits inverted.
pwr2 work through memory addresses in steps of powers of two.
pwr2inv like pwr2, but with the all relevant address bits inverted.
rev work through the address range with the bits in the address range reversed.
revinv like rev, but with all the relevant address bits inverted.
--vm-addr-mlock
attempt to mlock pages into memory causing more memory pressure by preventing pages from
swapped out.
--vm-addr-numa
assign memory mapped pages to randomly selected NUMA nodes. This is disabled for systems that
do not support NUMA.
--vm-addr-ops N
stop N workers after N bogo addressing passes.
Memory transfer between parent and child processes stressor (Linux)
--vm-rw N
start N workers that transfer memory to/from a parent/child using process_vm_writev(2) and
process_vm_readv(2). This is feature is only supported on Linux. Memory transfers are only
verified if the --verify option is enabled.
--vm-rw-bytes N
mmap N bytes per vm-rw worker, the default is 16 MB. One can specify the size as % of total
available memory or in units of Bytes, KBytes, MBytes and GBytes using the suffix b, k, m or
g.
--vm-rw-ops N
stop vm-rw workers after N memory read/writes.
Memory unmap from a child process stressor
--vm-segv N
start N workers that create a child process that unmaps its address space causing a SIGSEGV
on return from the unmap.
--vm-segv-ops N
stop after N bogo vm-segv SIGSEGV faults.
Vmsplice stressor (Linux)
--vm-splice N
move data from memory to /dev/null through a pipe without any copying between kernel address
space and user address space using vmsplice(2) and splice(2). This is only available for
Linux.
--vm-splice-bytes N
transfer N bytes per vmsplice call, the default is 64 K. One can specify the size as % of
total available memory or in units of Bytes, KBytes, MBytes and GBytes using the suffix b, k,
m or g.
--vm-splice-ops N
stop after N bogo vm-splice operations.
Virtual Memory Area (VMA) stressor
--vma N
start M workers that create pthreads to mmap, munmap, mlock, munlock, madvise, msync,
mprotect, mincore and access 32 pages in a randomly selected virtual memory address space.
This is designed to trip races on VMA page modifications. Every 15 seconds a different
virtual address space is randomly chosen.
--vma-ops N
stop the vma stressors after N successful memory mappings.
Vector neural network instructions stressor
--vnni N
start N workers that exercise vector neural network instructions (VNNI) used in convolutional
neural network loops. A 256 byte vector is operated upon using 8 bit multiply with 16 bit
summation, 16 bit multiply and 32 bit summation, and 8 bit summation. When processor features
allow, these operations using 512, 256 and 128 bit vector operations. Generic non-vectorized
code variants also provided (which may be vectorized by more advanced optimising compilers).
--vnni-intrinsic
just use the vnni methods that use intrinsic VNNI instructions and ignore the generic non-
vectorized methods.
--vnni-method N
select the VNNI method to be exercised, may be one of:
Method Description
all exercise all the following VNNI methods
vpaddb512 8 bit vector addition using 512 bit vector operations on 64 × 8 bit integers,
(x86 vpaddb)
vpaddb256 8 bit vector addition using 256 bit vector operations on 32 × 8 bit integers,
(x86 vpaddb)
vpaddb128 8 bit vector addition using 128 bit vectors operations on 32 × 8 bit integers,
(x86 vpaddb)
vpaddb 8 bit vector addition using 8 bit sequential addition (may be vectorized by the
compiler)
vpdpbusd512 8 bit vector multiplication of unsigned and signed 8 bit values followed by 16
bit summation using 512 bit vector operations on 64 × 8 bit integers, (x86
vpdpbusd)
vpdpbusd256 8 bit vector multiplication of unsigned and signed 8 bit values followed by 16
bit summation using 256 bit vector operations on 32 × 8 bit integers, (x86
vpdpbusd)
vpdpbusd128 8 bit vector multiplication of unsigned and signed 8 bit values followed by 16
bit summation using 128 bit vector operations on 32 × 8 bit integers, (x86
vpdpbusd)
vpdpbusd 8 bit vector multiplication of unsigned and signed 8 bit values followed by 16
bit summation using sequential operations (may be vectorized by the compiler)
vpdpwssd512 16 bit vector multiplication of unsigned and signed 16 bit values followed by 32
bit summation using 512 bit vector operations on 64 × 8 bit integers, (x86
vpdpwssd)
vpdpwssd256 16 bit vector multiplication of unsigned and signed 16 bit values followed by 32
bit summation using 256 bit vector operations on 64 × 8 bit integers, (x86
vpdpwssd)
vpdpwssd128 16 bit vector multiplication of unsigned and signed 16 bit values followed by 32
bit summation using 128 bit vector operations on 64 × 8 bit integers, (x86
vpdpwssd)
vpdpwssd 16 bit vector multiplication of unsigned and signed 16 bit values followed by 32
bit summation using sequential operations (may be vectorized by the compiler)
--vnni-ops N
stop after N bogo VNNI computation operations. 1 bogo-op is equivalent to 1024 convolution
loops operating on 256 bytes of data.
Pausing and resuming threads stressor
--wait N
start N workers that spawn off two children; one spins in a pause(2) loop, the other
continually stops and continues the first. The controlling process waits on the first child
to be resumed by the delivery of SIGCONT using waitpid(2) and waitid(2).
--wait-ops N
stop after N bogo wait operations.
CPU wait instruction stressor
--waitcpu N
start N workers that exercise processor wait instructions. For x86 these are pause, tpause
and umwait (when available) and nop. For ARM the yield instruction is used. For PPC64 the
yield, mdoio and mdooom instructions are used. For RISC-V the pause instruction is used. For
Loong64 the dbar instruction is used. For other architectures no-op instructions are used.
--waitcpu-ops N
stop after N bogo processor wait operations.
Watchdog stressor
--watchdog N
start N workers that exercising the /dev/watchdog watchdog interface by opening it, perform
various watchdog specific ioctl(2) commands on the device and close it. Before closing the
special watchdog magic close message is written to the device to try and force it to never
trip a watchdog reboot after the stressor has been run. Note that this stressor needs to be
run as root with the --pathological option and is only available on Linux.
--watchdog-ops N
stop after N bogo operations on the watchdog device.
Libc wide characterstring function stressor
--wcs N
start N workers that exercise various libc wide character string functions on random strings.
--wcs-method wcsfunc
select a specific libc wide character string function to stress. Available string functions
to stress are: all, wcscasecmp, wcscat, wcschr, wcscoll, wcscmp, wcscpy, wcslen, wcsncasecmp,
wcsncat, wcsncmp, wcsrchr and wcsxfrm. The `all' method is the default and will exercise all
the string methods.
--wcs-ops N
stop after N bogo wide character string operations.
scheduler workload stressor
--workload N
start N workers that exercise the scheduler with items of work that are started at random
times with random sleep delays between work items. By default a 100,000 microsecond slice of
time has 100 work items that start at random times during the slice. The work items by
default run for a quanta of 1000 microseconds scaled by the percentage work load (default of
30%). For a slice of S microseconds and a work item quanta duration of Q microseconds, S / Q
work items are executed per slice. For a work load of L percent, the run time per item is the
quanta Q × L / 100 microseconds. The --workload-threads option allows work items to be taken
from a queue and run concurrently if the scheduling run times overlap.
If a work item is already running when a new work item is scheduled to run then the new work
item is delayed and starts directly after the completion of the currently running work item
when running with the default of zero worker threads. This emulates bursty scheduled
compute, such as handling input packets where one may have lots of work items bunched
together or with random unpredictable delays between work items.
--workload-load L
specify the percentage run time load of each work item with respect to the run quanta
duration. Essentially the run duration of each work item is the quanta duration Q × L / 100.
--workload-method method
select the workload method. Each quanta of execution time is consumed using a tight spin-loop
executing a workload method. The available methods are described as follows:
Method Description
all randomly select any one of all the following methods:
fma perform multiply-add operations, on modern processors these may be
compiled into fused-multiply-add instructions.
getpid get the stressor's PID via getpid(2).
time get the current time via time(2).
inc64 increment a 64 bit integer.
memmove copy (move) a 1 MB buffer using memmove(3).
memread read from a 1 MB buffer using fast memory reads.
memset write to a 1 MB buffer using memset(3).
mcw64 compute 64 bit random numbers using a mwc random generator.
nop waste cycles using no-op instructions.
pause stop execution using CPU pause/yield or memory barrier instructions
where available.
random a random mix of all the workload methods, changing the workload method
on every spin-loop.
sqrt perform double precision floating point sqrt(3) and hypot(3) math
operations.
vecfp perform multiplication and addition on a vector of 64 double precision
floating point values.
--workload-sched [ batch | deadline | fifo | idle | other | rr ]
select scheduling policy. Note that deadline, fifo and rr require root privilege.
--workload-slice-us S
specify the duration of each scheduling slice in microseconds. The default is 100,000
microseconds (0.1 seconds).
--workload-quanta-us Q
specify the duration of each work item in microseconds. The default is 1000 microseconds (1
millisecond).
--workload-threads N
use N process threads to take scheduler work items of a workqueue and run the work item
(default is 2). When N is 0, no threads are used and the work items are run back-to-back
sequentially without using work queue. Using more than 2 threads allows work items to be
handled concurrently if enough idle processors are available.
--workload-dist [ cluster | even | poisson | random1 | random2 | random3 ]
specify the scheduling distribution of work items, the default is cluster. The distribution
methods are described as follows:
Method Description
cluster cluster 2/3 of the start times to try to start at the random time during the time
slice, with the other 1/3 of start times evenly randomly distributed using a single
random variable. The clustered start times causes a burst of items to be scheduled
in a bunch with no delays between each clustered work item.
even evenly distribute scheduling start times across the workload slice
poisson generate scheduling events that occur individually at random moments, but which tend
to occur at an average rate (known as a Poisson process).
random1 evenly randomly distribute scheduling start times using a single random variable.
random2 randomly distribute scheduling start times using a sum of two random variables, much
like throwing 2 dice.
random3 randomly distribute scheduling start times using a sum of three random variables,
much like throwing 3 dice.
--workload-ops N
stop the workload workers after N workload bogo-operations.
x86 cpuid stressor
--x86cpuid N
start N workers that exercise the x86 cpuid instruction with 18 different leaf types.
--x86cpuid-ops N
stop after N iterations that exercise the different cpuid leaf types.
x86-64 syscall stressor (Linux)
--x86syscall N
start N workers that repeatedly exercise the x86-64 syscall instruction to call the
getcpu(2), geteuid(2), getgid(2), getpid(2), gettimeofday(2) and time(2) system calls using
the Linux vsyscall handler. Only for Linux.
--x86syscall-func F
Instead of exercising the 6 syscall system calls, just call the syscall function F. The
function F must be one of getcpu, geteuid, getgid, getpid, gettimeofday or time.
--x86syscall-ops N
stop after N x86syscall system calls.
Extended file attributes stressor
--xattr N
start N workers that create, update and delete batches of extended attributes on a file.
--xattr-ops N
stop after N bogo extended attribute operations.
Yield scheduling stressor
-y N, --yield N
start N workers that call sched_yield(2). This stressor ensures that at least 2 child
processes per CPU exercise shield_yield(2) no matter how many workers are specified, thus
always ensuring rapid context switching.
--yield-ops N
stop yield stress workers after N sched_yield(2) bogo operations.
--yield-procs N
specify the number of child processes to run per stressor instance. The default is 2, range 1
to 65536.
--yield-sched [ batch | deadline | fifo | idle | other | rr ]
select scheduling policy. Note that deadline, fifo and rr require root privilege.
/dev/zero stressor
--zero N
start N workers that exercise /dev/zero with reads, lseeks, ioctls and mmaps. For just
/dev/zero read benchmarking use the --zero-read option.
--zero-ops N
stop zero stress workers after N /dev/zero bogo read operations.
--zero-read
just read /dev/zero with 4 K reads with no additional exercising on /dev/zero.
Zlib stressor
--zlib N
start N workers compressing and decompressing random data using zlib. Each worker has two
processes, one that compresses random data and pipes it to another process that decompresses
the data. This stressor exercises CPU, cache and memory.
--zlib-level L
specify the compression level (0..9), where 0 = no compression, 1 = fastest compression and 9
= best compression.
--zlib-mem-level L
specify the reserved compression state memory for zlib. Default is 8.
Value
1 minimum memory usage.
9 maximum memory usage.
--zlib-method method
specify the type of random data to send to the zlib library. By default, the data stream is
created from a random selection of the different data generation processes. However one can
specify just one method to be used if required. Available zlib data generation methods are
described as follows:
Method Description
00ff randomly distributed 0x00 and 0xFF values.
ascii01 randomly distributed ASCII 0 and 1 characters.
asciidigits randomly distributed ASCII digits in the range of 0 and 9.
bcd packed binary coded decimals, 0..99 packed into 2 4-bit nybbles.
binary 32 bit random numbers.
brown 8 bit brown noise (Brownian motion/Random Walk noise).
double double precision floating point numbers from sin(θ).
fixed data stream is repeated 0x04030201.
gcr random values as 4 × 4 bit data turned into 4 × 5 bit group coded recording
(GCR) patterns. Each 5 bit GCR value starts or ends with at most one zero bit
so that concatenated GCR codes have no more than two zero bits in a row.
gray 16 bit gray codes generated from an incrementing counter.
inc16 16 bit incrementing values starting from a random 16 bit value.
latin Random latin sentences from a sample of Lorem Ipsum text.
lehmer Fast random values generated using Lehmer's generator using a 128 bit multiply.
lfsr32 Values generated from a 32 bit Galois linear feedback shift register using the
polynomial x↑32 + x↑31 + x↑29 + x + 1. This generates a ring of 2↑32 − 1
unique values (all 32 bit values except for 0).
logmap Values generated from a logistical map of the equation Χn+1 = r × Χn × (1 − Χn)
where r > ∼ 3.56994567 to produce chaotic data. The values are scaled by a large
arbitrary value and the lower 8 bits of this value are compressed.
lrand48 Uniformly distributed pseudo-random 32 bit values generated from lrand48(3).
morse Morse code generated from random latin sentences from a sample of Lorem Ipsum
text.
nybble randomly distributed bytes in the range of 0x00 to 0x0f.
objcode object code selected from a random start point in the stress-ng text segment.
parity 7 bit binary data with 1 parity bit.
pink pink noise in the range 0..255 generated using the Gardner method with the
McCartney selection tree optimization. Pink noise is where the power spectral
density is inversely proportional to the frequency of the signal and hence is
slightly compressible.
random segments of the data stream are created by randomly calling the different data
generation methods.
rarely1 data that has a single 1 in every 32 bits, randomly located.
rarely0 data that has a single 0 in every 32 bits, randomly located.
rdrand generate random data using rdrand instruction (x86) or use 64 bit mwc psuedo-
random number generator for non-x86 systems.
ror32 generate a 32 bit random value, rotate it right 0 to 7 places and store the
rotated value for each of the rotations.
text random ASCII text.
utf8 random 8 bit data encoded to UTF-8.
zero all zeros, compresses very easily.
--zlib-ops N
stop after N bogo compression operations, each bogo compression operation is a compression of
64 K of random data at the highest compression level.
--zlib-strategy S
specifies the strategy to use when deflating data. This is used to tune the compression
algorithm. Default is 0.
Value
0 used for normal data (Z_DEFAULT_STRATEGY).
1 for data generated by a filter or predictor (Z_FILTERED)
2 forces huffman encoding (Z_HUFFMAN_ONLY).
3 Limit match distances to one run-length-encoding (Z_RLE).
4 prevents dynamic huffman codes (Z_FIXED).
--zlib-stream-bytes S
specify the amount of bytes to deflate until deflate should finish the block and return with
Z_STREAM_END. One can specify the size in units of Bytes, KBytes, MBytes and GBytes using the
suffix b, k, m or g. Default is 0 which creates and endless stream until stressor ends.
Value
0 creates an endless deflate stream until stressor stops.
n creates an stream of n bytes over and over again.
Each block will be closed with Z_STREAM_END.
--zlib-window-bits W
specify the window bits used to specify the history buffer size. The value is specified as
the base two logarithm of the buffer size (e.g. value 9 is 2↑9 = 512 bytes). Default is 15.
Value
-8-(-15) raw deflate format.
8-15 zlib format.
24-31 gzip format.
40-47 inflate auto format detection using zlib deflate format.
Zombie processes stressor
--zombie N
start N workers that create zombie processes. This will rapidly try to create a default of
8192 child processes that immediately die and wait in a zombie state until they are reaped.
Once the maximum number of processes is reached (or fork fails because one has reached the
maximum allowed number of children) the oldest child is reaped and a new process is then
created in a first-in first-out manner, and then repeated.
--zombie-max N
try to create as many as N zombie processes. This may not be reached if the system limit is
less than N.
--zombie-ops N
stop zombie stress workers after N bogo zombie operations.
EXAMPLES
stress-ng --vm 8 --vm-bytes 80% -t 1h
run 8 virtual memory stressors that combined use 80% of the available memory for 1 hour. Thus each
stressor uses 10% of the available memory.
stress-ng --vm 50% -t 1h
run virtual memory stressors on 50% of the number of online cpus for 1 hour.
stress-ng --vm 200% -t 1h
run virtual memory stressors on 2 times the number of online cpus for 1 hour.
stress-ng --cpu 4 --io 2 --vm 1 --vm-bytes 1G --timeout 60s
runs for 60 seconds with 4 cpu stressors, 2 io stressors and 1 vm stressor using 1 GB of virtual
memory.
stress-ng --iomix 2 --iomix-bytes 10% -t 10m
runs 2 instances of the mixed I/O stressors using a total of 10% of the available file system
space for 10 minutes. Each stressor will use 5% of the available file system space.
stress-ng --with cpu,matrix,vecmath,fp --seq 8 -t 1m
run 8 instances of cpu, matrix, vecmath and fp stressors sequentially one after another, for 1
minute per stressor.
stress-ng --with cpu,matrix,vecmath,fp --permute 5 -t 10s
run permutations of 5 instances of cpu, matrix, vecmath and fp stressors sequentially one after
another, for 10 seconds per permutation mix.
stress-ng --cyclic 1 --cyclic-dist 2500 --cyclic-method clock_ns --cyclic-prio 100 --cyclic-sleep 10000
--hdd 0 -t 1m
measures real time scheduling latencies created by the hdd stressor. This uses the high resolution
nanosecond clock to measure latencies during sleeps of 10,000 nanoseconds. At the end of 1 minute
of stressing, the latency distribution with 2500 ns intervals will be displayed. NOTE: this must
be run with the CAP_SYS_NICE capability to enable the real time scheduling to get accurate
measurements.
stress-ng --cpu 8 --cpu-ops 800000
runs 8 cpu stressors and stops after 800000 bogo operations.
stress-ng --sequential 2 --timeout 2m --metrics
run 2 simultaneous instances of all the stressors sequentially one by one, each for 2 minutes and
summarise with performance metrics at the end.
stress-ng --cpu 4 --cpu-method fft --cpu-ops 10000 --metrics-brief
run 4 FFT cpu stressors, stop after 10000 bogo operations and produce a summary just for the FFT
results.
stress-ng --cpu -1 --cpu-method all -t 1h --cpu-load 90
run cpu stressors on all online CPUs working through all the available CPU stressors for 1 hour,
loading the CPUs at 90% load capacity.
stress-ng --cpu 0 --cpu-method all -t 20m
run cpu stressors on all configured CPUs working through all the available CPU stressors for 20
minutes
stress-ng --all 4 --timeout 5m
run 4 instances of all the stressors for 5 minutes.
stress-ng --random 64
run 64 stressors that are randomly chosen from all the available stressors.
stress-ng --cpu 64 --cpu-method all --verify -t 10m --metrics-brief
run 64 instances of all the different cpu stressors and verify that the computations are correct
for 10 minutes with a bogo operations summary at the end.
stress-ng --sequential -1 -t 10m
run all the stressors one by one for 10 minutes, with the number of instances of each stressor
matching the number of online CPUs.
stress-ng --sequential 8 --class io -t 5m --times
run all the stressors in the io class one by one for 5 minutes each, with 8 instances of each
stressor running concurrently and show overall time utilisation statistics at the end of the run.
stress-ng --class io\?
show all the stressors in the io class
stress-ng --all -1 --maximize --aggressive
run all the stressors (1 instance of each per online CPU) simultaneously, maximize the settings
(memory sizes, file allocations, etc.) and select the most demanding/aggressive options.
stress-ng --all 8 --with cpu,hash,nop,vm --timeout 1m
run 8 instances of cpu, hash, nop and vm stressors altogether for 1 minute.
stress-ng --seq 8 --with cpu,hash,nop,vm --timeout 1m --progress
run 8 instances of cpu, hash, nop and vm stressors one after another for 1 minute each and show
the run progress.
stress-ng --random 32 -x numa,hdd,key
run 32 randomly selected stressors and exclude the numa, hdd and key stressors
stress-ng --sequential 4 --class vm --exclude bigheap,brk,stack
run 4 instances of the VM stressors one after another, excluding the bigheap, brk and stack
stressors
stress-ng --taskset 0,2-3 --cpu 3
run 3 instances of the CPU stressor and pin them to CPUs 0, 2 and 3.
stress-ng --taskset odd --cpu 32
run 32 instances of the CPU stressor and pin them to odd numbered CPUs.
EXIT STATUS
Status Description
0 Success.
1 Error; incorrect user options or a fatal resource issue in the stress-ng stressor harness
(for example, out of memory).
2 One or more stressors failed.
3 One or more stressors failed to initialise because of lack of resources, for example ENOMEM
(no memory), ENOSPC (no space on file system) or a missing or unimplemented system call.
4 One or more stressors were not implemented on a specific architecture or operating system.
5 A stressor has been killed by an unexpected signal.
6 A stressor exited by exit(2) which was not expected and timing metrics could not be
gathered.
7 The bogo ops metrics maybe untrustworthy. This is most likely to occur when a stress test is
terminated during the update of a bogo-ops counter such as when it has been OOM killed. A
less likely reason is that the counter ready indicator has been corrupted.
BUGS
File bug reports at: https://github.com/ColinIanKing/stress-ng/issues - please note that no support will
be provided if stress-ng is packaged without this manual.
SEE ALSO
cpuburn(1), perf(1), stress(1), taskset(1)
https://github.com/ColinIanKing/stress-ng/blob/master/README.md
AUTHOR
stress-ng was written by Colin Ian King <colin.i.king@gmail.com> and is a clean room re-implementation
and extension of the original stress tool by Amos Waterland. Thanks also to the many contributors to
stress-ng. The README.md file in the source contains a full list of the contributors.
NOTES
Sending a SIGALRM, SIGINT or SIGHUP to stress-ng causes it to terminate all the stressor processes and
ensures temporary files and shared memory segments are removed cleanly.
Sending a SIGUSR2 to stress-ng will dump out the current load average and memory statistics.
Note that the stress-ng cpu, io, vm and hdd tests are different implementations of the original stress
tests and hence may produce different stress characteristics.
The bogo operations metrics may change with each release because of bug fixes to the code, new features,
compiler optimisations, changes in support libraries or system call performance.
COPYRIGHT
Copyright © 2013-2021 Canonical Ltd, Copyright © 2021-2025 Colin Ian King.
This is free software; see the source for copying conditions. There is NO warranty; not even for
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
26 June 2025 STRESS-NG(1)