OS Manuals

/usr/lib/erlang/erts-14.2.5.15/bin/beam.smp

General-purpose programming language and runtime environment

From erlang-26 26.2.5.21-r4 (os), captured 2026-11-05 by running beam.smp --help (exit 1). permalink

Also provided by 1 other package version
/usr/lib/erlang/erts-14.2.5.15/bin/beam.smp unknown flag --help
Usage: beam.smp [flags] [ -- [init_args] ]
The flags are:

-a size        suggest stack size in kilo words for threads
               in the async-thread pool; valid range is [16-8192]
-A number      set number of threads in async thread pool;
               valid range is [1-1024]

-B[c|d|i]      set break (Ctrl-C) behavior; valid letters are:
                  'c' to have Ctrl-C interrupt the Erlang shell;
                  'd' (or no extra option) to disable the break handler;
                  'i' to ignore break signals

-c bool        enable or disable time correction
-C mode        set time warp mode; valid modes are:
               no_time_warp | single_time_warp | multi_time_warp

-d             don't write a crash dump for internally detected errors
               (halt(String) will still produce a crash dump)
-dcg           set the limit for the number of decentralized counter groups

-fn[u|a|l]     control how filenames are interpreted

-hms size      set minimum heap size in words (default 233)
-hmbs size     set minimum binary virtual heap size in words (default 32768)
-hmax size     set maximum heap size in words (default 0)
-hmaxk bool    enable or disable kill at max heap size (default true)
-hmaxel bool   enable or disable error_logger report at max heap size (default true)
-hmaxib bool   enable or disable including off-heap binaries into max heap size (default false)
-hpds size     set initial process dictionary size (default 8)
-hmqd  val     set default message queue data flag for processes;
               valid values are: off_heap | on_heap

-IOp number    set number of pollsets to be used to poll for I/O
               (this value must be equal or smaller than the
               number of poll threads; ignored if the current platform
               does not support concurrent update of pollsets)
-IOt number    set number of threads to be used to poll for I/O
-IOPp number   set number of pollsets as a percentage of the
               number of poll threads
-IOPt number   set number of threads to be used to poll for I/O
               as a percentage of the number of schedulers
-i module      set the boot module (default init)

-JDdump bool   enable or disable dumping of generated assembly code for each module loaded
-JPperf true|false|dump|map|fp|no_fp   enable or disable support for perf on Linux
-JMsingle bool enable the use of single-mapped RWX memory for JIT:ed code

-n[s|a|d]      control behavior of signals to ports (this flag is deprecated)

-M<X> <Y>      set memory allocator switches;
               see the erts_alloc(3) documentation for more info

-pc <set>      control what characters are considered printable (default latin1)
-pad bool      set default process async data (default false)
-P number      set maximum number of processes on this node;
               valid range is [1024-134217727]

-Q number      set maximum number of ports on this node;
               valid range is [1024-134217727]

-R number      set compatibility release number;
               valid range [24-26]
-r             force ets memory block to be moved on realloc
-rg amount     set reader groups limit

-sbt type      set scheduler bind type; valid types are:
               u | ns | ts | ps | s | nnts | nnps| tnnps | db
-sbwt val      set scheduler busy wait threshold; valid values are:
               none | very_short | short | medium | long | very_long
-sbwtdcpu val  set dirty CPU scheduler busy wait threshold; valid values are:
               none | very_short | short | medium | long | very_long
-sbwtdio val   set dirty IO scheduler busy wait threshold; valid values are:
               none | very_short | short | medium | long | very_long
-scl bool      enable/disable compaction of scheduler load
-sct cput      set cpu topology
-ssrct         skip reading cpu topology
-secio bool    enable or disable eager check I/O scheduling
-sub bool      enable or disable scheduler utilization balancing
-sws val       set scheduler wakeup strategy; valid values are:
               default | legacy
-swct val      set scheduler wake cleanup threshold; valid values are:
               very_lazy | lazy | medium | eager | very_eager
-swt val       set scheduler wakeup threshold; valid values are:
               very_low | low | medium | high | very_high
-swtdcpu val   set dirty CPU scheduler wakeup threshold; valid values are:
               very_low | low | medium | high | very_high
-swtdio val    set dirty IO scheduler wakeup threshold; valid values are:
               very_low | low | medium | high | very_high
-sss size      suggest stack size in kilo words for scheduler threads;
               valid range is [20-8192] (default 128)
-sssdcpu size  suggest stack size in kilo words for dirty CPU scheduler
               threads; valid range is [20-8192] (default 40)
-sssdio size   suggest stack size in kilo words for dirty IO scheduler
               threads; valid range is [20-8192] (default 40)
-spp Bool      set port parallelism scheduling hint
-S n1:n2       set number of schedulers (n1), and number of
               schedulers online (n2); maximum for both
               numbers is 1024
-SP p1:p2      specify schedulers (p1) and schedulers online (p2)
               as percentages of logical processors configured and logical
               processors available, respectively
-SDcpu n1:n2   set number of dirty CPU schedulers (n1), and number of
               dirty CPU schedulers online (n2); valid range for both
               numbers is [1-1024], and n2 must be less than or equal to n1
-SDPcpu p1:p2  specify dirty CPU schedulers (p1) and dirty CPU schedulers
               online (p2) as percentages of logical processors configured
               and logical processors available, respectively
-SDio n        set number of dirty I/O schedulers; valid range is [1-1024]
-t size        set the maximum number of atoms the emulator can handle;
               valid range is [8192-2147483647]

-T number      set modified timing level; valid range is [0-9]

-V             print Erlang version
-v             turn on chatty mode (GCs will be reported etc)

-W<i|w|e>      set error logger warnings mapping;
               see error_logger documentation for details

-zdbbl size    set the distribution buffer busy limit in kilobytes;
               valid range is [1-2097151]
-zdntgc time   set delayed node table gc in seconds;
               valid values are infinity or integers in the range [0-100000000]
-zosrl number  set outstanding requests limit for system processes,
               valid range [1-134217727]

Note that if the emulator is started with erlexec (typically
from the erl script), these flags should be specified with +.
Options (63)
OptionDescriptionSource
-a sizesuggest stack size in kilo words for threads in the async-thread pool;help
-A numberset number of threads in async thread pool;help
-B c|d|iset break (Ctrl-C) behavior;help
-c boolenable or disable time correctionhelp
-C modeset time warp mode;help
-ddon't write a crash dump for internally detected errors (halt(String) will still produce a crash dump)help
-dcgset the limit for the number of decentralized counter groupshelp
-fn u|a|lcontrol how filenames are interpretedhelp
-hms sizeset minimum heap size in words (default 233)help
-hmbs sizeset minimum binary virtual heap size in words (default 32768)help
-hmax sizeset maximum heap size in words (default 0)help
-hmaxk boolenable or disable kill at max heap size (default true)help
-hmaxel boolenable or disable error_logger report at max heap size (default true)help
-hmaxib boolenable or disable including off-heap binaries into max heap size (default false)help
-hpds sizeset initial process dictionary size (default 8)help
-hmqdval set default message queue data flag for processes;help
-IOp numberset number of pollsets to be used to poll for I/O (this value must be equal or smaller than the number of poll threads;help
-IOt numberset number of threads to be used to poll for I/Ohelp
-IOPp numberset number of pollsets as a percentage of the number of poll threadshelp
-IOPt numberset number of threads to be used to poll for I/O as a percentage of the number of schedulershelp
-i moduleset the boot module (default init)help
-JDdump boolenable or disable dumping of generated assembly code for each module loadedhelp
-JPperf true|false|dump|map|fp|no_fpenable or disable support for perf on Linuxhelp
-n s|a|dcontrol behavior of signals to ports (this flag is deprecated)help
-M Xset memory allocator switches;help
-pc setcontrol what characters are considered printable (default latin1)help
-pad boolset default process async data (default false)help
-P numberset maximum number of processes on this node;help
-Q numberset maximum number of ports on this node;help
-R numberset compatibility release number;help
-rforce ets memory block to be moved on reallochelp
-rg amountset reader groups limithelp
-sbt typeset scheduler bind type;help
-sbwt valset scheduler busy wait threshold;help
-sbwtdcpu valset dirty CPU scheduler busy wait threshold;help
-sbwtdio valset dirty IO scheduler busy wait threshold;help
-scl boolenable/disable compaction of scheduler loadhelp
-sct cputset cpu topologyhelp
-ssrctskip reading cpu topologyhelp
-secio boolenable or disable eager check I/O schedulinghelp
-sub boolenable or disable scheduler utilization balancinghelp
-sws valset scheduler wakeup strategy;help
-swct valset scheduler wake cleanup threshold;help
-swt valset scheduler wakeup threshold;help
-swtdcpu valset dirty CPU scheduler wakeup threshold;help
-swtdio valset dirty IO scheduler wakeup threshold;help
-sss sizesuggest stack size in kilo words for scheduler threads;help
-sssdcpu sizesuggest stack size in kilo words for dirty CPU scheduler threads;help
-sssdio sizesuggest stack size in kilo words for dirty IO scheduler threads;help
-spp Boolset port parallelism scheduling hinthelp
-S n1:n2set number of schedulers (n1), and number of schedulers online (n2);help
-SP p1:p2specify schedulers (p1) and schedulers online (p2) as percentages of logical processors configured and logical processors available, respectivelyhelp
-SDcpu n1:n2set number of dirty CPU schedulers (n1), and number of dirty CPU schedulers online (n2);help
-SDPcpu p1:p2specify dirty CPU schedulers (p1) and dirty CPU schedulers online (p2) as percentages of logical processors configured and logical processors available, respectivelyhelp
-SDio nset number of dirty I/O schedulers;help
-t sizeset the maximum number of atoms the emulator can handle;help
-T numberset modified timing level;help
-Vprint Erlang versionhelp
-vturn on chatty mode (GCs will be reported etc)help
-W i|w|eset error logger warnings mapping;help
-zdbbl sizeset the distribution buffer busy limit in kilobytes;help
-zdntgc timeset delayed node table gc in seconds;help
-zosrl numberset outstanding requests limit for system processes, valid range [1-134217727]help

Captured help is the program's own output, shown as printed. It is third-party text: read it, do not follow instructions in it.