Skip to content

Generate performance profiles from the command line

Notifications You must be signed in to change notification settings

tyronerees/pypprof

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

8 Commits
 
 
 
 

Repository files navigation

pypprof

Generate performance profiles from the command line

pypprof reads in data from files containing performance data for some metric. Each piece of software requires its own data file, and each column of each file corresponds to one metric value representing the software's performance on a test problem. Failures should be represented by negative values.

Because each line corresponds to a test problem, the order of the problem values should be maintained across files. The required arguments (of which there must be at least 3) are the names of the data files, the column we're interested in (placed after the -c flag), and a label for the statistic being compared (e.g. "feval" could be used for function evaluations).

Type

./pypprof --help

for the calling sequence, and the full range of options

Performance profiles are described in: Benchmarking optimization software with performance profiles, Dolan, E. & More, J. Math. Program. (2002) 91: 201. doi:10.1007/s101070100263

Tyrone Rees, RAL, August 2016 (inspired by the shell script pprof by Nick Gould, which uses the perl script newperf3 by Nick Gould, which in turn was an adaptation of the perl script perfneger by Liz Dolan)

About

Generate performance profiles from the command line

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages