1. mike fc
  2. numpy-benchmark

Source

numpy-benchmark /

Filename Size Date modified Message
181 B
1.1 KB
656 B
3.7 KB
5.2 KB

Micro-benchmarks for some numpy functions

  • Benchmarking to compare numpy implementations
  • Micro-benchmarks using timeit
  • Currently, most benchmarks are for in-place operations as I'm wanting to test algorithm time, and not conflate it with memory allocation time.

Usage

  • ./benchmark.py [numpy library name] (defaults to numpy)
  • writes timings to a file of tab-separated values (TSV)
  • Graph with benchmark.Rmd in Rstudio
  • just set the names of the TSV files to compare.

Output