A program to run benchmarks
runlim [ options ...] command [ arguments ...]
run is a tool that can be used to run and control benchmarks. It executes a given command with (optional) arguments, samples resource usage during the run, and kills the process (and its child processes) if a certain time and/or space limit is exhausted.
Every 100 milliseconds, runlim takes a sample of the program's resource utilization, and logs status information to stderr every second. Optionally, the status can be logged to a file.
Multi-threaded programs can be limited by setting a wall clock timeout. runlim follows the time accumulation scheme of GNU time for multi-threaded programs and programs that spawn multiple child-processes: time spent in each thread/child is summed up, unless you are only interested in walk clock time.
runlim accepts the following options:
-h, --help
Show summary of options.
--version
Show version of program.
-o FILE, --output-file=FILE
Overwrite or create FILE for output logging.
-s NUM, --space-limit=NUM
Set space limit to NUM megabytes.
-t NUM, --time-limit=NUM
Set time limit to NUM seconds.
-r NUM, --real-time-limit=NUM
Set real time limit to NUM seconds.
-k, --kill
Propagate signals.
time(1), timelimit(1), timeout(1), time(7).
runlim was written by Armin Biere and Toni Jussila.
This manual page was written by Thomas Krennwallner <[email protected]>, for the Debian project (and may be used by others).