lean4-htt/tests/bench
2019-05-29 17:36:42 +02:00
..
.gitignore
arith_eval.ml
binarytrees.ghc-6.hs
binarytrees.lean
binarytrees.ocaml-2.ml
compile.sh
default.nix chore(tests/bench/default.nix): remove compilers of retired categories 2019-05-29 17:36:42 +02:00
deriv.hs
deriv.lean
deriv.ml
deriv.sml
expr_const_folding.hs
expr_const_folding.lean
expr_const_folding.ml
expr_const_folding.sml
expr_const_folding.swift
frontend_test.lean
ghc-gc.py
lean-gc.py
leanpkg.path
Makefile feat(tests/bench/Makefile): update HTML report 2019-05-29 17:13:55 +02:00
mlkit-gc.py
nixpkgs.nix
ocaml-gc.py
perf.py
qsort.hs
qsort.lean
qsort.ml
qsort.sml
rbmap.hs
rbmap.lean
rbmap.library.lean
rbmap.ml
rbmap.sml
rbmap2.lean
rbmap3.lean
rbmap4.lean
rbmap_checkpoint.hs
rbmap_checkpoint.lean
rbmap_checkpoint.ml
rbmap_checkpoint.sml
rbmap_checkpoint2.lean
rbmap_checkpoint2.sml
README.md chore(tests/bench): add readme 2019-05-29 17:32:35 +02:00
report.py
run.sh
temci.yaml
unionfind1.lean
unionfind2.lean

Lean Benchmark Suite

This folder contains multiple small Lean programs for benchmarking and comparing their performance to other Lean configurations and functional compilers using the temci benchmarking tool.

We recommend using Nix for building/obtaining all Lean variants and used compilers in a reproducible way. After installing Nix, running the benchmarks is as easy as

nix-shell --run make

This will record 50 runs for each benchmark configuration (this can be changed with min_runs in temci.yaml), generate results in report_lean.csv and report_cross.csv, and print them to stdout in a tabulated format. It will also generate HTML reports in report/ comparing the time-based benchmarks.

In order to reduce noise in the benchmarking data, you may instead want to try calling make inside a temci shell:

nix-shell --run "temci short shell --sudo --preset usable --cpuset_active make"

Using root powers, this will temporarily configure your machine similarly to the LLVM benchmarking recommendations and move all your other processes to a single CPU core.