HigherOrderCO / HVM

A massively parallel, optimal functional runtime in Rust
https://higherorderco.com
Apache License 2.0
10.44k stars 397 forks source link

add bench workflow #348

Closed tjjfvi closed 4 months ago

HigherOrderBot commented 4 months ago

Perf run for 62fe15d:

Your branch is up to date with 'origin/main'.
compiled
========

file            runtime         main            (local)       
==============================================================
sort_bitonic    c                        5.58s           8.25s
                cuda                     error           error
--------------------------------------------------------------
sum_rec         c                        1.42s           2.75s
                cuda                     error           error
--------------------------------------------------------------
sum_tree        c                        0.12s           0.28s
                cuda                     error           error
--------------------------------------------------------------
tuples          c                        3.39s           4.80s
                cuda                     error           error
--------------------------------------------------------------

interpreted
===========

file            runtime         main            (local)       
==============================================================
sort_bitonic    c                        5.86s           9.30s
                cuda                     error           error
                rust                    24.33s          39.53s
--------------------------------------------------------------
sum_rec         c                        2.57s           2.52s
                cuda                     error           error
                rust                    13.82s          13.68s
--------------------------------------------------------------
sum_tree        c                        0.26s           0.38s
                cuda                     error           error
                rust                     0.87s           1.46s
--------------------------------------------------------------
tuples          c                        2.47s           3.81s
                cuda                     error           error
                rust                     3.73s           6.33s
--------------------------------------------------------------
HigherOrderBot commented 4 months ago

Perf run for 4870145:

hi i'm a table
HigherOrderBot commented 4 months ago

Perf run for dd108f8:

compiled
========

file            runtime         main            (local)       
==============================================================
sort_bitonic    c                        3.45s           5.74s
                cuda                     0.24s           2.25s
--------------------------------------------------------------
sum_rec         c                        1.44s           1.63s
                cuda                     0.14s           0.36s
--------------------------------------------------------------
sum_tree        c                        0.35s           0.15s
                cuda                     0.09s           0.18s
--------------------------------------------------------------
tuples          c                        3.13s           4.36s
                cuda                   timeout         timeout
--------------------------------------------------------------

interpreted
===========

file            runtime         main            (local)       
==============================================================
sort_bitonic    c                        4.48s           5.19s
                cuda                     0.23s           2.19s
                rust                   timeout         timeout
--------------------------------------------------------------
sum_rec         c                        1.69s           2.40s
                cuda                     0.14s           0.35s
                rust                    13.92s          13.87s
--------------------------------------------------------------
sum_tree        c                        0.17s           0.36s
                cuda                     0.08s           0.19s
                rust                     0.87s           0.87s
--------------------------------------------------------------
tuples          c                        2.48s           2.48s
                cuda                   timeout         timeout
                rust                     3.61s           3.66s
--------------------------------------------------------------