Skip to content

Instantly share code, notes, and snippets.

@abadams
Created April 5, 2018 00:40
Show Gist options
  • Save abadams/d77665c791678d97e07803aca91908cc to your computer and use it in GitHub Desktop.
Save abadams/d77665c791678d97e07803aca91908cc to your computer and use it in GitHub Desktop.
fast_synchronization windows performance
ubuntu-on-windows-10 in branch
performance_inner_loop_parallel 71.7ms
performance_parallel_performance 0.020
bilateral grid 16.89 / 10.9
blur 7.00
camera pipe 7.162 / 32.739
conv layer 7.09 / 6.03
interpolate 36.815
lens blur 8.10 / 36.44
local laplacian 74.23 / 76.138
nl_means 209.138 / 177.876
resize:
planar lanczos float32 0.12 time: 4.523400 ms
packed lanczos float32 0.12 time: 11.939900 ms
planar box float32 4.00 time: 0.552600 ms
packed box float32 4.00 time: 0.636300 ms
planar box float32 0.50 time: 2.495700 ms
packed box float32 0.50 time: 3.709500 ms
planar box uint16 4.00 time: 0.352500 ms
packed box uint16 4.00 time: 0.513200 ms
planar box uint16 0.50 time: 1.440600 ms
packed box uint16 0.50 time: 1.991500 ms
planar box uint8 4.00 time: 0.373200 ms
packed box uint8 4.00 time: 0.505400 ms
planar box uint8 0.50 time: 1.314700 ms
packed box uint8 0.50 time: 1.738800 ms
planar linear float32 4.00 time: 0.557600 ms
packed linear float32 4.00 time: 1.124800 ms
planar linear float32 0.50 time: 2.709200 ms
packed linear float32 0.50 time: 4.144200 ms
planar linear uint16 4.00 time: 0.439600 ms
packed linear uint16 4.00 time: 0.602700 ms
planar linear uint16 0.50 time: 2.084100 ms
packed linear uint16 0.50 time: 3.118500 ms
planar linear uint8 4.00 time: 0.457200 ms
packed linear uint8 4.00 time: 0.602600 ms
planar linear uint8 0.50 time: 2.004500 ms
packed linear uint8 0.50 time: 2.974700 ms
planar cubic float32 4.00 time: 0.799100 ms
packed cubic float32 4.00 time: 1.837500 ms
planar cubic float32 0.50 time: 3.777600 ms
packed cubic float32 0.50 time: 5.729600 ms
planar cubic uint16 4.00 time: 0.845000 ms
packed cubic uint16 4.00 time: 0.999300 ms
planar cubic uint16 0.50 time: 3.520800 ms
packed cubic uint16 0.50 time: 5.758500 ms
planar cubic uint8 4.00 time: 0.852600 ms
packed cubic uint8 4.00 time: 0.981300 ms
planar cubic uint8 0.50 time: 3.361600 ms
packed cubic uint8 0.50 time: 5.630100 ms
planar lanczos float32 4.00 time: 1.153700 ms
packed lanczos float32 4.00 time: 2.040300 ms
planar lanczos float32 0.50 time: 5.419700 ms
packed lanczos float32 0.50 time: 8.366900 ms
planar lanczos uint16 4.00 time: 1.199700 ms
packed lanczos uint16 4.00 time: 1.286900 ms
planar lanczos uint16 0.50 time: 5.231500 ms
packed lanczos uint16 0.50 time: 8.621800 ms
planar lanczos uint8 4.00 time: 1.212800 ms
packed lanczos uint8 4.00 time: 1.249400 ms
planar lanczos uint8 0.50 time: 5.032800 ms
packed lanczos uint8 0.50 time: 8.495800 ms
ubuntu-on-windows-10 in master
performance_inner_loop_parallel 290ms
performance_parallel_performance 0.020
bilateral grid 17 / 10.8
blur 6.99
camera pipe 7.333 / 32.575
conv layer 7.1 / 6.08
interpolate 37.145
lens blur 8.07 / 36.17
local laplacian 74.06 / 74.215
nl_means 208.645 / 177.378
resize:
planar lanczos float32 0.12 time: 4.561100 ms
packed lanczos float32 0.12 time: 12.106000 ms
planar box float32 4.00 time: 0.544800 ms
packed box float32 4.00 time: 0.641600 ms
planar box float32 0.50 time: 2.492100 ms
packed box float32 0.50 time: 3.697400 ms
planar box uint16 4.00 time: 0.345300 ms
packed box uint16 4.00 time: 0.508700 ms
planar box uint16 0.50 time: 1.450400 ms
packed box uint16 0.50 time: 2.044900 ms
planar box uint8 4.00 time: 0.365400 ms
packed box uint8 4.00 time: 0.506900 ms
planar box uint8 0.50 time: 1.319500 ms
packed box uint8 0.50 time: 1.738200 ms
planar linear float32 4.00 time: 0.557100 ms
packed linear float32 4.00 time: 1.150300 ms
planar linear float32 0.50 time: 2.701000 ms
packed linear float32 0.50 time: 4.122500 ms
planar linear uint16 4.00 time: 0.448000 ms
packed linear uint16 4.00 time: 0.598000 ms
planar linear uint16 0.50 time: 2.116800 ms
packed linear uint16 0.50 time: 3.149900 ms
planar linear uint8 4.00 time: 0.449000 ms
packed linear uint8 4.00 time: 0.608100 ms
planar linear uint8 0.50 time: 1.997200 ms
packed linear uint8 0.50 time: 2.972700 ms
planar cubic float32 4.00 time: 0.800400 ms
packed cubic float32 4.00 time: 1.938700 ms
planar cubic float32 0.50 time: 3.806200 ms
packed cubic float32 0.50 time: 5.759700 ms
planar cubic uint16 4.00 time: 0.862500 ms
packed cubic uint16 4.00 time: 0.998500 ms
planar cubic uint16 0.50 time: 3.478700 ms
packed cubic uint16 0.50 time: 5.766800 ms
planar cubic uint8 4.00 time: 0.848000 ms
packed cubic uint8 4.00 time: 0.970900 ms
planar cubic uint8 0.50 time: 3.376100 ms
packed cubic uint8 0.50 time: 5.632800 ms
planar lanczos float32 4.00 time: 1.156600 ms
packed lanczos float32 4.00 time: 2.004500 ms
planar lanczos float32 0.50 time: 5.459200 ms
packed lanczos float32 0.50 time: 8.438700 ms
planar lanczos uint16 4.00 time: 1.217000 ms
packed lanczos uint16 4.00 time: 1.335500 ms
planar lanczos uint16 0.50 time: 5.281900 ms
packed lanczos uint16 0.50 time: 8.658000 ms
planar lanczos uint8 4.00 time: 1.209300 ms
packed lanczos uint8 4.00 time: 1.310200 ms
planar lanczos uint8 0.50 time: 5.134700 ms
packed lanczos uint8 0.50 time: 8.517700 ms
mingw in branch
performance_inner_loop_parallel 100.8
performance_parallel_performance 0.0183
bilateral grid 11.48 / 11.09
blur 8.066
camera pipe 6.730 / 24.955
conv layer 9.03 / 6.28
interpolate 21.63
lens blur 12.97 / 26.91
local laplacian 64.17 / 85.33
nl_means 215.2 / 176.175
resize:
planar lanczos float32 0.12 time: 5.817683 ms
packed lanczos float32 0.12 time: 12.455127 ms
planar box float32 4.00 time: 0.681491 ms
packed box float32 4.00 time: 0.673034 ms
planar box float32 0.50 time: 2.538076 ms
packed box float32 0.50 time: 3.742603 ms
planar box uint16 4.00 time: 0.375134 ms
packed box uint16 4.00 time: 0.559929 ms
planar box uint16 0.50 time: 1.480813 ms
packed box uint16 0.50 time: 2.124925 ms
planar box uint8 4.00 time: 0.405307 ms
packed box uint8 4.00 time: 0.539132 ms
planar box uint8 0.50 time: 1.334596 ms
packed box uint8 0.50 time: 1.837245 ms
planar linear float32 4.00 time: 0.692502 ms
packed linear float32 4.00 time: 0.709645 ms
planar linear float32 0.50 time: 2.749748 ms
packed linear float32 0.50 time: 4.199724 ms
planar linear uint16 4.00 time: 0.494779 ms
packed linear uint16 4.00 time: 0.653004 ms
planar linear uint16 0.50 time: 2.154613 ms
packed linear uint16 0.50 time: 3.237936 ms
planar linear uint8 4.00 time: 0.547589 ms
packed linear uint8 4.00 time: 0.607322 ms
planar linear uint8 0.50 time: 2.032949 ms
packed linear uint8 0.50 time: 3.035870 ms
planar cubic float32 4.00 time: 0.898273 ms
packed cubic float32 4.00 time: 0.985114 ms
planar cubic float32 0.50 time: 3.922161 ms
packed cubic float32 0.50 time: 5.859379 ms
planar cubic uint16 4.00 time: 0.909616 ms
packed cubic uint16 4.00 time: 1.035675 ms
planar cubic uint16 0.50 time: 3.614629 ms
packed cubic uint16 0.50 time: 5.862419 ms
planar cubic uint8 4.00 time: 0.921267 ms
packed cubic uint8 4.00 time: 1.020831 ms
planar cubic uint8 0.50 time: 3.533103 ms
packed cubic uint8 0.50 time: 5.686796 ms
planar lanczos float32 4.00 time: 1.830219 ms
packed lanczos float32 4.00 time: 1.668137 ms
planar lanczos float32 0.50 time: 6.512995 ms
packed lanczos float32 0.50 time: 9.073503 ms
planar lanczos uint16 4.00 time: 1.867188 ms
packed lanczos uint16 4.00 time: 1.752142 ms
planar lanczos uint16 0.50 time: 6.575948 ms
packed lanczos uint16 0.50 time: 9.471044 ms
planar lanczos uint8 4.00 time: 1.895190 ms
packed lanczos uint8 4.00 time: 1.778457 ms
planar lanczos uint8 0.50 time: 6.164228 ms
packed lanczos uint8 0.50 time: 9.207635 ms
mingw in master
performance_inner_loop_parallel 170
performance_parallel_performance 0.0183
bilateral grid 11.48 / 11.07
blur 8.124
camera pipe 6.876 / 26.008
conv layer 9.95 / 6.26
interpolate 21.76
lens blur 13.13 / 26.66
local laplacian 63.91 / 82.25
nl_means 215.5 / 177.327
resize:
planar lanczos float32 0.12 time: 5.928641 ms
packed lanczos float32 0.12 time: 12.268134 ms
planar box float32 4.00 time: 0.616699 ms
packed box float32 4.00 time: 0.719047 ms
planar box float32 0.50 time: 2.522977 ms
packed box float32 0.50 time: 3.817768 ms
planar box uint16 4.00 time: 0.368006 ms
packed box uint16 4.00 time: 0.544549 ms
planar box uint16 0.50 time: 1.493715 ms
packed box uint16 0.50 time: 2.183049 ms
planar box uint8 4.00 time: 0.391128 ms
packed box uint8 4.00 time: 0.520635 ms
planar box uint8 0.50 time: 1.322026 ms
packed box uint8 0.50 time: 1.851961 ms
planar linear float32 4.00 time: 0.647306 ms
packed linear float32 4.00 time: 0.681593 ms
planar linear float32 0.50 time: 2.744894 ms
packed linear float32 0.50 time: 4.362342 ms
planar linear uint16 4.00 time: 0.470687 ms
packed linear uint16 4.00 time: 0.658778 ms
planar linear uint16 0.50 time: 2.144087 ms
packed linear uint16 0.50 time: 3.274676 ms
planar linear uint8 4.00 time: 0.517952 ms
packed linear uint8 4.00 time: 0.634430 ms
planar linear uint8 0.50 time: 2.041278 ms
packed linear uint8 0.50 time: 3.042411 ms
planar cubic float32 4.00 time: 0.877502 ms
packed cubic float32 4.00 time: 1.004071 ms
planar cubic float32 0.50 time: 3.975583 ms
packed cubic float32 0.50 time: 5.910502 ms
planar cubic uint16 4.00 time: 0.920322 ms
packed cubic uint16 4.00 time: 1.050314 ms
planar cubic uint16 0.50 time: 3.614578 ms
packed cubic uint16 0.50 time: 5.885438 ms
planar cubic uint8 4.00 time: 0.937107 ms
packed cubic uint8 4.00 time: 1.093645 ms
planar cubic uint8 0.50 time: 3.521989 ms
packed cubic uint8 0.50 time: 5.686975 ms
planar lanczos float32 4.00 time: 1.885635 ms
packed lanczos float32 4.00 time: 1.751963 ms
planar lanczos float32 0.50 time: 6.678373 ms
packed lanczos float32 0.50 time: 9.192817 ms
planar lanczos uint16 4.00 time: 1.892788 ms
packed lanczos uint16 4.00 time: 1.796086 ms
planar lanczos uint16 0.50 time: 6.518642 ms
packed lanczos uint16 0.50 time: 9.492326 ms
planar lanczos uint8 4.00 time: 1.932415 ms
packed lanczos uint8 4.00 time: 1.827077 ms
planar lanczos uint8 0.50 time: 6.212208 ms
packed lanczos uint8 0.50 time: 9.246776 ms
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment