Software for benchmarking programs using the GNU Time command, with extensive benchmarking options and unified behaviour across platforms (including output units). Inspired by the benchmarking script in Phylign developed by Leandro Lima.
$ if [[ $(uname) == "Darwin" ]]; then brew install gnu-time; fi
$ conda install -y -c bioconda -c conda-forge galitime
$ galitime --log time.log ls
$ cat time.log
real_s user_s sys_s percent_cpu max_ram_kb exit_code fs_inputs fs_outputs real_s_alt command
0.03 0.00 0.00 10% 2904 0 0 0 0.076661 lsGalitime has no dependencies beyond Python 3. While it used to require gtime on OS X, this dependency is no longer required and the default time command is used.
conda install -y -c bioconda -c conda-forge galitimeInstall the Galitime Python package:
pip install -U galitime$ galitime -h
usage: galitime [-h] --log LOG [--experiment EXPERIMENT] [-v] command
Benchmark a command.
positional arguments:
command The command to be benchmarked
options:
-h, --help show this help message and exit
--log LOG Path to the log file with benchmark statistics (if the directory doesn't exist, it will be created).
--experiment EXPERIMENT
Name of the experiment (to be attached to the output)
-v show program's version number and exit
-
experiment- Name of the experiment, if provided via-n -
run- Number of run in case of multiple repetitions -
real_s- Real time in seconds (wall clock time) -
real_s_py- Python-measured real time in seconds (slightly higher thanreal_s) -
user_s- User CPU time in seconds (user mode, excluding system calls) -
sys_s- System CPU time in seconds (kernel mode) -
percent_cpu- CPU usage percentage -
max_ram_kb- Maximum RAM usage in kilobytes -
exit_code- Exit code of the command (0 if everything ok, -1 if command failed) -
fs_inputs- File system read read operations count -
fs_outputs- File system write operations count -
command- Command executed, with tabs replaced by spaces
Please use Github issues.
See Releases.
