Add uint8 support for interpolate for CPU images by NicolasHug · Pull Request #90771 · pytorch/pytorch

NicolasHug · 2022-12-13T16:13:33Z

Joint work with @vfdev-5

This PR introduces native uint8 support for interpolate(), for bilinear ~~and bicubic~~ modes for CPU images (mode=nearest[_exact] was already supported ).

On a typical torchvision training job on ImageNet, the speedup are ~4X when AVX2 is supported, comparing the uint8 native (this PR) vs torchvision's current Resize():

AA = antialias
float = uint8->float->interpolate()->round()->clamp()->uint8 (what Resize() currently does)

input_size         output_size channels_last AA    mode       num_threads  speed-up float vs uint8 (this PR)
(1, 3, 270, 268) -> (224, 224)     True    True    bilinear   num_threads=1   4X    2.6ms vs 0.7ms
(1, 3, 270, 268) -> (224, 224)     True    False   bilinear   num_threads=1   2.1X  1.3ms vs 0.6ms
(1, 3, 270, 268) -> (224, 224)     False   True    bilinear   num_threads=1   3X    2.1ms vs 0.7ms
(1, 3, 270, 268) -> (224, 224)     False   False   bilinear   num_threads=1   4X    2.4ms vs 0.6ms

(Note: we removed bicubic support for now)
(1, 3, 270, 268) -> (224, 224)     True    True    bicubic    num_threads=1   4X    2.9ms vs 0.7ms
(1, 3, 270, 268) -> (224, 224)     True    False   bicubic    num_threads=1   5X    3.1ms vs 0.7ms
(1, 3, 270, 268) -> (224, 224)     False   True    bicubic    num_threads=1   3X    2.4ms vs 0.7ms
(1, 3, 270, 268) -> (224, 224)     False   False   bicubic    num_threads=1   4X    2.8ms vs 0.7ms

There is still room for further speed-ups (see TODOs in the code).

More benchmark details

with AVX2 support - speedups typically range from 1.5X to 10X. A few edge-cases are slower, worth investigating why.

Details

AA = antialias
float = uint8->float->interpolate()->round()->clamp()->uint8 (what Resize() currently does)

input_size         output_size channels_last AA    mode       num_threads  speed-up float vs uint8 (this PR)
(1, 3, 64, 64) -> (224, 224)       True    True    bilinear   num_threads=1   5X    1.1ms vs 0.2ms
(1, 3, 64, 64) -> (224, 224)       True    False   bilinear   num_threads=1   5X    1.2ms vs 0.2ms
(1, 3, 64, 64) -> (224, 224)       False   True    bilinear   num_threads=1   2.8X  0.6ms vs 0.2ms
(1, 3, 64, 64) -> (224, 224)       False   False   bilinear   num_threads=1   7X    1.6ms vs 0.2ms
(1, 3, 64, 64) -> (224, 224)       True    True    bicubic    num_threads=1   5X    1.2ms vs 0.2ms
(1, 3, 64, 64) -> (224, 224)       True    False   bicubic    num_threads=1   12X   2.9ms vs 0.2ms
(1, 3, 64, 64) -> (224, 224)       False   True    bicubic    num_threads=1   3X    0.8ms vs 0.2ms
(1, 3, 64, 64) -> (224, 224)       False   False   bicubic    num_threads=1   7X    1.8ms vs 0.2ms

(1, 3, 64, 64) -> (224, 224)       True    True    bilinear   num_threads=2   2.6X  0.6ms vs 0.2ms
(1, 3, 64, 64) -> (224, 224)       True    False   bilinear   num_threads=2   2.8X  0.6ms vs 0.2ms
(1, 3, 64, 64) -> (224, 224)       False   True    bilinear   num_threads=2   1.7X  0.4ms vs 0.2ms
(1, 3, 64, 64) -> (224, 224)       False   False   bilinear   num_threads=2   1.4X  0.3ms vs 0.2ms
(1, 3, 64, 64) -> (224, 224)       True    True    bicubic    num_threads=2   2.7X  0.7ms vs 0.2ms
(1, 3, 64, 64) -> (224, 224)       True    False   bicubic    num_threads=2   7X    1.6ms vs 0.2ms
(1, 3, 64, 64) -> (224, 224)       False   True    bicubic    num_threads=2   1.8X  0.4ms vs 0.2ms
(1, 3, 64, 64) -> (224, 224)       False   False   bicubic    num_threads=2   4X    1.0ms vs 0.2ms

(1, 3, 224, 224) -> (270, 268)     True    True    bilinear   num_threads=1   4X    2.5ms vs 0.6ms
(1, 3, 224, 224) -> (270, 268)     True    False   bilinear   num_threads=1   3.0X  1.8ms vs 0.6ms
(1, 3, 224, 224) -> (270, 268)     False   True    bilinear   num_threads=1   3X    1.8ms vs 0.6ms
(1, 3, 224, 224) -> (270, 268)     False   False   bilinear   num_threads=1   4X    2.3ms vs 0.6ms
(1, 3, 224, 224) -> (270, 268)     True    True    bicubic    num_threads=1   4X    2.7ms vs 0.6ms
(1, 3, 224, 224) -> (270, 268)     True    False   bicubic    num_threads=1   7X    4.3ms vs 0.6ms
(1, 3, 224, 224) -> (270, 268)     False   True    bicubic    num_threads=1   3X    2.1ms vs 0.6ms
(1, 3, 224, 224) -> (270, 268)     False   False   bicubic    num_threads=1   4X    2.6ms vs 0.6ms

(1, 3, 224, 224) -> (270, 268)     True    True    bilinear   num_threads=2   2.7X  1.6ms vs 0.6ms
(1, 3, 224, 224) -> (270, 268)     True    False   bilinear   num_threads=2   2.6X  1.5ms vs 0.6ms
(1, 3, 224, 224) -> (270, 268)     False   True    bilinear   num_threads=2   2.1X  1.2ms vs 0.6ms
(1, 3, 224, 224) -> (270, 268)     False   False   bilinear   num_threads=2   1.6X  0.9ms vs 0.6ms
(1, 3, 224, 224) -> (270, 268)     True    True    bicubic    num_threads=2   2.8X  1.7ms vs 0.6ms
(1, 3, 224, 224) -> (270, 268)     True    False   bicubic    num_threads=2   5X    2.8ms vs 0.6ms
(1, 3, 224, 224) -> (270, 268)     False   True    bicubic    num_threads=2   2.3X  1.4ms vs 0.6ms
(1, 3, 224, 224) -> (270, 268)     False   False   bicubic    num_threads=2   3X    1.9ms vs 0.6ms

(1, 3, 256, 256) -> (1024, 1024)   True    True    bilinear   num_threads=1   4X    26.6ms vs 6.7ms
(1, 3, 256, 256) -> (1024, 1024)   True    False   bilinear   num_threads=1   4X    23.9ms vs 6.8ms
(1, 3, 256, 256) -> (1024, 1024)   False   True    bilinear   num_threads=1   2.5X  16.8ms vs 6.8ms
(1, 3, 256, 256) -> (1024, 1024)   False   False   bilinear   num_threads=1   5X    33.1ms vs 6.8ms
(1, 3, 256, 256) -> (1024, 1024)   True    True    bicubic    num_threads=1   4X    25.9ms vs 7.3ms
(1, 3, 256, 256) -> (1024, 1024)   True    False   bicubic    num_threads=1   8X    59.6ms vs 7.3ms
(1, 3, 256, 256) -> (1024, 1024)   False   True    bicubic    num_threads=1   1.9X  14.3ms vs 7.4ms
(1, 3, 256, 256) -> (1024, 1024)   False   False   bicubic    num_threads=1   5X    35.4ms vs 7.3ms

(1, 3, 256, 256) -> (1024, 1024)   True    True    bilinear   num_threads=2   2.0X  13.6ms vs 6.8ms
(1, 3, 256, 256) -> (1024, 1024)   True    False   bilinear   num_threads=2   2.2X  14.8ms vs 6.7ms
(1, 3, 256, 256) -> (1024, 1024)   False   True    bilinear   num_threads=2   1.3X  8.8ms vs 6.9ms
(1, 3, 256, 256) -> (1024, 1024)   False   False   bilinear   num_threads=2   1.2X  8.4ms vs 6.8ms
(1, 3, 256, 256) -> (1024, 1024)   True    True    bicubic    num_threads=2   1.8X  12.8ms vs 7.3ms
(1, 3, 256, 256) -> (1024, 1024)   True    False   bicubic    num_threads=2   4X    32.1ms vs 7.2ms
(1, 3, 256, 256) -> (1024, 1024)   False   True    bicubic    num_threads=2   1.4X  10.1ms vs 7.3ms
(1, 3, 256, 256) -> (1024, 1024)   False   False   bicubic    num_threads=2   2.9X  20.9ms vs 7.3ms

(1, 3, 224, 224) -> (64, 64)       True    True    bilinear   num_threads=1   1.4X  0.5ms vs 0.3ms
(1, 3, 224, 224) -> (64, 64)       True    False   bilinear   num_threads=1   0.7X  0.2ms vs 0.3ms
(1, 3, 224, 224) -> (64, 64)       False   True    bilinear   num_threads=1   1.3X  0.4ms vs 0.3ms
(1, 3, 224, 224) -> (64, 64)       False   False   bilinear   num_threads=1   1.4X  0.4ms vs 0.3ms
(1, 3, 224, 224) -> (64, 64)       True    True    bicubic    num_threads=1   2.1X  0.7ms vs 0.3ms
(1, 3, 224, 224) -> (64, 64)       True    False   bicubic    num_threads=1   1.3X  0.4ms vs 0.3ms
(1, 3, 224, 224) -> (64, 64)       False   True    bicubic    num_threads=1   1.9X  0.6ms vs 0.3ms
(1, 3, 224, 224) -> (64, 64)       False   False   bicubic    num_threads=1   1.0X  0.3ms vs 0.3ms

(1, 3, 224, 224) -> (64, 64)       True    True    bilinear   num_threads=2   1.0X  0.3ms vs 0.3ms
(1, 3, 224, 224) -> (64, 64)       True    False   bilinear   num_threads=2   0.6X  0.2ms vs 0.3ms
(1, 3, 224, 224) -> (64, 64)       False   True    bilinear   num_threads=2   0.8X  0.3ms vs 0.3ms
(1, 3, 224, 224) -> (64, 64)       False   False   bilinear   num_threads=2   1.4X  0.4ms vs 0.3ms
(1, 3, 224, 224) -> (64, 64)       True    True    bicubic    num_threads=2   1.4X  0.5ms vs 0.3ms
(1, 3, 224, 224) -> (64, 64)       True    False   bicubic    num_threads=2   1.2X  0.4ms vs 0.3ms
(1, 3, 224, 224) -> (64, 64)       False   True    bicubic    num_threads=2   1.2X  0.4ms vs 0.4ms
(1, 3, 224, 224) -> (64, 64)       False   False   bicubic    num_threads=2   0.9X  0.3ms vs 0.3ms

(1, 3, 270, 268) -> (224, 224)     True    True    bilinear   num_threads=1   4X    2.6ms vs 0.7ms
(1, 3, 270, 268) -> (224, 224)     True    False   bilinear   num_threads=1   2.1X  1.3ms vs 0.6ms
(1, 3, 270, 268) -> (224, 224)     False   True    bilinear   num_threads=1   3X    2.1ms vs 0.7ms
(1, 3, 270, 268) -> (224, 224)     False   False   bilinear   num_threads=1   4X    2.4ms vs 0.6ms
(1, 3, 270, 268) -> (224, 224)     True    True    bicubic    num_threads=1   4X    2.9ms vs 0.7ms
(1, 3, 270, 268) -> (224, 224)     True    False   bicubic    num_threads=1   5X    3.1ms vs 0.7ms
(1, 3, 270, 268) -> (224, 224)     False   True    bicubic    num_threads=1   3X    2.4ms vs 0.7ms
(1, 3, 270, 268) -> (224, 224)     False   False   bicubic    num_threads=1   4X    2.8ms vs 0.7ms

(1, 3, 270, 268) -> (224, 224)     True    True    bilinear   num_threads=2   1.5X  1.0ms vs 0.7ms
(1, 3, 270, 268) -> (224, 224)     True    False   bilinear   num_threads=2   1.2X  0.8ms vs 0.6ms
(1, 3, 270, 268) -> (224, 224)     False   True    bilinear   num_threads=2   2.3X  1.5ms vs 0.7ms
(1, 3, 270, 268) -> (224, 224)     False   False   bilinear   num_threads=2   1.9X  1.2ms vs 0.6ms
(1, 3, 270, 268) -> (224, 224)     True    True    bicubic    num_threads=2   1.6X  1.2ms vs 0.7ms
(1, 3, 270, 268) -> (224, 224)     True    False   bicubic    num_threads=2   4X    2.4ms vs 0.7ms
(1, 3, 270, 268) -> (224, 224)     False   True    bicubic    num_threads=2   2.4X  1.6ms vs 0.7ms
(1, 3, 270, 268) -> (224, 224)     False   False   bicubic    num_threads=2   2.8X  1.8ms vs 0.6ms

(1, 3, 1024, 1024) -> (256, 256)   True    True    bilinear   num_threads=1   2.1X  12.8ms vs 6.1ms
(1, 3, 1024, 1024) -> (256, 256)   True    False   bilinear   num_threads=1   0.6X  3.8ms vs 5.9ms
(1, 3, 1024, 1024) -> (256, 256)   False   True    bilinear   num_threads=1   1.2X  7.1ms vs 6.1ms
(1, 3, 1024, 1024) -> (256, 256)   False   False   bilinear   num_threads=1   1.9X  11.0ms vs 5.9ms
(1, 3, 1024, 1024) -> (256, 256)   True    True    bicubic    num_threads=1   2.0X  12.6ms vs 6.4ms
(1, 3, 1024, 1024) -> (256, 256)   True    False   bicubic    num_threads=1   1.0X  6.1ms vs 6.0ms
(1, 3, 1024, 1024) -> (256, 256)   False   True    bicubic    num_threads=1   1.8X  11.3ms vs 6.4ms
(1, 3, 1024, 1024) -> (256, 256)   False   False   bicubic    num_threads=1   0.8X  4.6ms vs 6.0ms

(1, 3, 1024, 1024) -> (256, 256)   True    True    bilinear   num_threads=2   1.6X  9.3ms vs 6.0ms
(1, 3, 1024, 1024) -> (256, 256)   True    False   bilinear   num_threads=2   0.3X  2.0ms vs 5.8ms
(1, 3, 1024, 1024) -> (256, 256)   False   True    bilinear   num_threads=2   1.2X  7.2ms vs 6.0ms
(1, 3, 1024, 1024) -> (256, 256)   False   False   bilinear   num_threads=2   0.3X  1.6ms vs 5.8ms
(1, 3, 1024, 1024) -> (256, 256)   True    True    bicubic    num_threads=2   1.1X  7.1ms vs 6.5ms
(1, 3, 1024, 1024) -> (256, 256)   True    False   bicubic    num_threads=2   0.6X  3.3ms vs 5.9ms
(1, 3, 1024, 1024) -> (256, 256)   False   True    bicubic    num_threads=2   0.9X  5.9ms vs 6.3ms
(1, 3, 1024, 1024) -> (256, 256)   False   False   bicubic    num_threads=2   0.4X  2.4ms vs 5.9ms

without AVX2 support - no significant speed-up, but there are various possible improvements (see TODOs)

Details

AA = antialias
float = uint8->float->interpolate()->round()->clamp()->uint8 (what Resize() currently does)

input_size         output_size channels_last AA    mode       num_threads  speed-up float vs uint8 (this PR)
(1, 3, 64, 64) -> (224, 224)       True    True    bilinear   num_threads=1   0.9X  1.5ms vs 1.6ms
(1, 3, 64, 64) -> (224, 224)       True    False   bilinear   num_threads=1   0.9X  1.5ms vs 1.6ms
(1, 3, 64, 64) -> (224, 224)       False   True    bilinear   num_threads=1   0.8X  0.9ms vs 1.1ms
(1, 3, 64, 64) -> (224, 224)       False   False   bilinear   num_threads=1   1.5X  1.7ms vs 1.1ms
(1, 3, 64, 64) -> (224, 224)       True    True    bicubic    num_threads=1   0.9X  1.6ms vs 1.8ms
(1, 3, 64, 64) -> (224, 224)       True    False   bicubic    num_threads=1   2.1X  3.9ms vs 1.9ms
(1, 3, 64, 64) -> (224, 224)       False   True    bicubic    num_threads=1   0.8X  1.1ms vs 1.4ms
(1, 3, 64, 64) -> (224, 224)       False   False   bicubic    num_threads=1   1.7X  2.4ms vs 1.5ms

(1, 3, 64, 64) -> (224, 224)       True    True    bilinear   num_threads=2   0.9X  0.8ms vs 0.8ms
(1, 3, 64, 64) -> (224, 224)       True    False   bilinear   num_threads=2   0.9X  0.8ms vs 0.8ms
(1, 3, 64, 64) -> (224, 224)       False   True    bilinear   num_threads=2   0.9X  0.5ms vs 0.6ms
(1, 3, 64, 64) -> (224, 224)       False   False   bilinear   num_threads=2   0.7X  0.5ms vs 0.7ms
(1, 3, 64, 64) -> (224, 224)       True    True    bicubic    num_threads=2   0.9X  0.9ms vs 1.0ms
(1, 3, 64, 64) -> (224, 224)       True    False   bicubic    num_threads=2   2.1X  2.0ms vs 1.0ms
(1, 3, 64, 64) -> (224, 224)       False   True    bicubic    num_threads=2   0.8X  0.6ms vs 0.8ms
(1, 3, 64, 64) -> (224, 224)       False   False   bicubic    num_threads=2   1.7X  1.3ms vs 0.8ms

(1, 3, 224, 224) -> (270, 268)     True    True    bilinear   num_threads=1   1.0X  3.0ms vs 3.0ms
(1, 3, 224, 224) -> (270, 268)     True    False   bilinear   num_threads=1   1.0X  2.8ms vs 2.9ms
(1, 3, 224, 224) -> (270, 268)     False   True    bilinear   num_threads=1   1.0X  2.3ms vs 2.2ms
(1, 3, 224, 224) -> (270, 268)     False   False   bilinear   num_threads=1   1.4X  3.3ms vs 2.3ms
(1, 3, 224, 224) -> (270, 268)     True    True    bicubic    num_threads=1   1.0X  3.5ms vs 3.5ms
(1, 3, 224, 224) -> (270, 268)     True    False   bicubic    num_threads=1   1.7X  6.1ms vs 3.5ms
(1, 3, 224, 224) -> (270, 268)     False   True    bicubic    num_threads=1   0.9X  2.6ms vs 2.9ms
(1, 3, 224, 224) -> (270, 268)     False   False   bicubic    num_threads=1   1.4X  4.2ms vs 2.9ms

(1, 3, 224, 224) -> (270, 268)     True    True    bilinear   num_threads=2   1.0X  1.7ms vs 1.7ms
(1, 3, 224, 224) -> (270, 268)     True    False   bilinear   num_threads=2   0.9X  1.6ms vs 1.8ms
(1, 3, 224, 224) -> (270, 268)     False   True    bilinear   num_threads=2   0.9X  1.3ms vs 1.4ms
(1, 3, 224, 224) -> (270, 268)     False   False   bilinear   num_threads=2   0.7X  1.1ms vs 1.6ms
(1, 3, 224, 224) -> (270, 268)     True    True    bicubic    num_threads=2   1.0X  2.0ms vs 2.0ms
(1, 3, 224, 224) -> (270, 268)     True    False   bicubic    num_threads=2   1.7X  3.2ms vs 1.9ms
(1, 3, 224, 224) -> (270, 268)     False   True    bicubic    num_threads=2   0.8X  1.5ms vs 1.9ms
(1, 3, 224, 224) -> (270, 268)     False   False   bicubic    num_threads=2   1.2X  2.3ms vs 1.9ms

(1, 3, 256, 256) -> (1024, 1024)   True    True    bilinear   num_threads=1   1.1X  34.7ms vs 32.4ms
(1, 3, 256, 256) -> (1024, 1024)   True    False   bilinear   num_threads=1   1.0X  31.2ms vs 32.4ms
(1, 3, 256, 256) -> (1024, 1024)   False   True    bilinear   num_threads=1   1.0X  23.5ms vs 22.7ms
(1, 3, 256, 256) -> (1024, 1024)   False   False   bilinear   num_threads=1   1.9X  42.5ms vs 22.7ms
(1, 3, 256, 256) -> (1024, 1024)   True    True    bicubic    num_threads=1   0.9X  33.9ms vs 37.4ms
(1, 3, 256, 256) -> (1024, 1024)   True    False   bicubic    num_threads=1   2.2X  84.0ms vs 37.5ms
(1, 3, 256, 256) -> (1024, 1024)   False   True    bicubic    num_threads=1   1.0X  28.4ms vs 28.8ms
(1, 3, 256, 256) -> (1024, 1024)   False   False   bicubic    num_threads=1   2.0X  56.7ms vs 28.8ms

(1, 3, 256, 256) -> (1024, 1024)   True    True    bilinear   num_threads=2   1.1X  17.5ms vs 16.4ms
(1, 3, 256, 256) -> (1024, 1024)   True    False   bilinear   num_threads=2   1.1X  17.7ms vs 16.4ms
(1, 3, 256, 256) -> (1024, 1024)   False   True    bilinear   num_threads=2   0.8X  8.8ms vs 11.4ms
(1, 3, 256, 256) -> (1024, 1024)   False   False   bilinear   num_threads=2   1.0X  11.1ms vs 11.4ms
(1, 3, 256, 256) -> (1024, 1024)   True    True    bicubic    num_threads=2   1.1X  19.9ms vs 18.8ms
(1, 3, 256, 256) -> (1024, 1024)   True    False   bicubic    num_threads=2   2.3X  42.5ms vs 18.7ms
(1, 3, 256, 256) -> (1024, 1024)   False   True    bicubic    num_threads=2   1.0X  14.1ms vs 14.5ms
(1, 3, 256, 256) -> (1024, 1024)   False   False   bicubic    num_threads=2   2.0X  28.4ms vs 14.5ms

(1, 3, 224, 224) -> (64, 64)       True    True    bilinear   num_threads=1   1.0X  0.6ms vs 0.6ms
(1, 3, 224, 224) -> (64, 64)       True    False   bilinear   num_threads=1   0.7X  0.3ms vs 0.4ms
(1, 3, 224, 224) -> (64, 64)       False   True    bilinear   num_threads=1   0.9X  0.5ms vs 0.6ms
(1, 3, 224, 224) -> (64, 64)       False   False   bilinear   num_threads=1   1.7X  0.6ms vs 0.4ms
(1, 3, 224, 224) -> (64, 64)       True    True    bicubic    num_threads=1   1.0X  0.8ms vs 0.8ms
(1, 3, 224, 224) -> (64, 64)       True    False   bicubic    num_threads=1   1.1X  0.5ms vs 0.5ms
(1, 3, 224, 224) -> (64, 64)       False   True    bicubic    num_threads=1   0.9X  0.7ms vs 0.8ms
(1, 3, 224, 224) -> (64, 64)       False   False   bicubic    num_threads=1   0.9X  0.4ms vs 0.4ms

(1, 3, 224, 224) -> (64, 64)       True    True    bilinear   num_threads=2   1.0X  0.4ms vs 0.4ms
(1, 3, 224, 224) -> (64, 64)       True    False   bilinear   num_threads=2   0.8X  0.2ms vs 0.3ms
(1, 3, 224, 224) -> (64, 64)       False   True    bilinear   num_threads=2   0.9X  0.3ms vs 0.3ms
(1, 3, 224, 224) -> (64, 64)       False   False   bilinear   num_threads=2   1.3X  0.3ms vs 0.2ms
(1, 3, 224, 224) -> (64, 64)       True    True    bicubic    num_threads=2   1.0X  0.5ms vs 0.5ms
(1, 3, 224, 224) -> (64, 64)       True    False   bicubic    num_threads=2   1.3X  0.4ms vs 0.3ms
(1, 3, 224, 224) -> (64, 64)       False   True    bicubic    num_threads=2   0.9X  0.5ms vs 0.5ms
(1, 3, 224, 224) -> (64, 64)       False   False   bicubic    num_threads=2   1.2X  0.3ms vs 0.3ms

(1, 3, 270, 268) -> (224, 224)     True    True    bilinear   num_threads=1   0.8X  2.1ms vs 2.5ms
(1, 3, 270, 268) -> (224, 224)     True    False   bilinear   num_threads=1   0.7X  1.6ms vs 2.4ms
(1, 3, 270, 268) -> (224, 224)     False   True    bilinear   num_threads=1   1.2X  2.4ms vs 2.1ms
(1, 3, 270, 268) -> (224, 224)     False   False   bilinear   num_threads=1   1.3X  2.6ms vs 2.0ms
(1, 3, 270, 268) -> (224, 224)     True    True    bicubic    num_threads=1   1.1X  3.4ms vs 3.0ms
(1, 3, 270, 268) -> (224, 224)     True    False   bicubic    num_threads=1   1.7X  4.8ms vs 2.8ms
(1, 3, 270, 268) -> (224, 224)     False   True    bicubic    num_threads=1   1.1X  2.9ms vs 2.7ms
(1, 3, 270, 268) -> (224, 224)     False   False   bicubic    num_threads=1   1.4X  3.5ms vs 2.4ms

(1, 3, 270, 268) -> (224, 224)     True    True    bilinear   num_threads=2   0.9X  1.2ms vs 1.3ms
(1, 3, 270, 268) -> (224, 224)     True    False   bilinear   num_threads=2   1.3X  1.6ms vs 1.2ms
(1, 3, 270, 268) -> (224, 224)     False   True    bilinear   num_threads=2   0.8X  0.9ms vs 1.1ms
(1, 3, 270, 268) -> (224, 224)     False   False   bilinear   num_threads=2   1.3X  1.3ms vs 1.0ms
(1, 3, 270, 268) -> (224, 224)     True    True    bicubic    num_threads=2   1.4X  2.2ms vs 1.6ms
(1, 3, 270, 268) -> (224, 224)     True    False   bicubic    num_threads=2   1.9X  2.8ms vs 1.5ms
(1, 3, 270, 268) -> (224, 224)     False   True    bicubic    num_threads=2   0.8X  1.1ms vs 1.4ms
(1, 3, 270, 268) -> (224, 224)     False   False   bicubic    num_threads=2   1.7X  2.1ms vs 1.3ms

(1, 3, 1024, 1024) -> (256, 256)   True    True    bilinear   num_threads=1   1.0X  10.0ms vs 9.9ms
(1, 3, 1024, 1024) -> (256, 256)   True    False   bilinear   num_threads=1   0.7X  4.6ms vs 6.2ms
(1, 3, 1024, 1024) -> (256, 256)   False   True    bilinear   num_threads=1   0.9X  9.1ms vs 9.8ms
(1, 3, 1024, 1024) -> (256, 256)   False   False   bilinear   num_threads=1   1.7X  9.4ms vs 5.7ms
(1, 3, 1024, 1024) -> (256, 256)   True    True    bicubic    num_threads=1   1.0X  15.2ms vs 14.8ms
(1, 3, 1024, 1024) -> (256, 256)   True    False   bicubic    num_threads=1   1.0X  7.6ms vs 7.5ms
(1, 3, 1024, 1024) -> (256, 256)   False   True    bicubic    num_threads=1   0.9X  13.3ms vs 14.4ms
(1, 3, 1024, 1024) -> (256, 256)   False   False   bicubic    num_threads=1   0.8X  5.9ms vs 7.0ms

(1, 3, 1024, 1024) -> (256, 256)   True    True    bilinear   num_threads=2   1.2X  6.0ms vs 5.2ms
(1, 3, 1024, 1024) -> (256, 256)   True    False   bilinear   num_threads=2   0.7X  2.3ms vs 3.2ms
(1, 3, 1024, 1024) -> (256, 256)   False   True    bilinear   num_threads=2   1.0X  4.8ms vs 5.0ms
(1, 3, 1024, 1024) -> (256, 256)   False   False   bilinear   num_threads=2   0.7X  1.9ms vs 2.9ms
(1, 3, 1024, 1024) -> (256, 256)   True    True    bicubic    num_threads=2   1.6X  12.3ms vs 7.5ms
(1, 3, 1024, 1024) -> (256, 256)   True    False   bicubic    num_threads=2   1.0X  3.9ms vs 3.9ms
(1, 3, 1024, 1024) -> (256, 256)   False   True    bicubic    num_threads=2   1.0X  7.0ms vs 7.3ms
(1, 3, 1024, 1024) -> (256, 256)   False   False   bicubic    num_threads=2   0.9X  3.0ms vs 3.5ms

Benchmark code

Details

import operator_benchmark as op_bench
import torch

"""Microbenchmarks for interpolate operator."""


class InterpolateBenchmark(op_bench.TorchBenchmarkBase):
    def init(self, input_size, output_size, channels_last=False, mode='linear', antialias=False, dtype=torch.float):

        input_image = torch.randint(0, 256, size=input_size, dtype=torch.uint8, device='cpu')
                                    
        if channels_last:
            input_image = input_image.contiguous(memory_format=torch.channels_last)

        self.inputs = {
            "input_image": input_image,
            "output_size": output_size,
            "mode": mode,
            "antialias": antialias,
            "dtype":dtype,
        }

        self.set_module_name("interpolate")

    def forward(self, input_image, output_size, mode, antialias, dtype):
        if dtype == torch.float:
            input_image = input_image.float()

        out = torch.nn.functional.interpolate(input_image, size=output_size, mode=mode, align_corners=False, antialias=antialias)
        if dtype == torch.float:
            out = out.round().clamp(min=0, max=256).to(torch.uint8)


def make_config():
    sizes = (
        ((224, 224), (64, 64)),
        ((270, 268), (224, 224)),
        ((256, 256), (1024, 1024)),
    )

    attrs = []
    for (HW1, HW2) in sizes:
        attrs.append([(1, 3, *HW1), HW2])  # 3 channels
        # attrs.append([(1, 1, *HW1), HW2])  # 1 channel

        attrs.append([(1, 3, *HW2), HW1])  # 3 channels
        # attrs.append([(1, 1, *HW2), HW1])  # 1 channel

    config = op_bench.config_list(
        attr_names=["input_size", "output_size"],
        attrs=attrs,
        cross_product_configs={
            'channels_last': [True, False],
            'mode': ["bilinear", "bicubic"],
            'antialias': [True, False],
            # 'dtype': [torch.float, torch.uint8]
            # 'dtype': [torch.uint8]
            'dtype': [torch.float]
        },
        tags=["short"],
    )

    return config

config = make_config()
op_bench.generate_pt_test(config, InterpolateBenchmark)


if __name__ == "__main__":
    op_bench.benchmark_runner.main()

import re
import argparse

parser = argparse.ArgumentParser()
parser.add_argument("f1", nargs="?", default="main")
parser.add_argument("f2", nargs="?", default="new")
args = parser.parse_args()

with open(args.f1) as f:
    main = f.readlines()
with open(args.f2) as f:
    new = f.readlines()

out = []

for main_line, new_line in zip(main, new):
    # num_threads=1  # TODO: remove
    if main_line.startswith("num_threads="):
        num_threads = int(main_line.split("=")[-1])
    if main_line.startswith("# Input"):
        deets = f"{main_line.strip()}, {num_threads=}"
    if main_line.startswith("Forward"):
        main_time = float(main_line.split()[-1])
        new_time = float(new_line.split()[-1])
        ratio = main_time / new_time
        fmt = ".1f" if ratio < 3 else ".0f"
        improv = f"{ratio:{fmt}}X"
        time_fmt = ",.3f" if new_time < 100 else ",.1f"
        deets = deets.strip().replace("# Input: ", "")
        deets = deets.replace(": ", "=")
        deets = deets.replace("input_size=", "")
        deets = deets.replace(", output_size=", " -> ")
        deets = deets.replace("dtype=torch.", "")
        deets = deets.replace("mode=", "")
        deets = deets.replace("antialias=", "")
        deets = deets.replace("channels_last=", "")
        # deets = deets.replace("channels_last=True, ", "")
        split = deets.split(",")

        # size = ','.join(split[:-3])
        # mode, dtype, threads = split[-3:]
        # deets = f"{size:<30} {mode:<15} {dtype:<10} {threads:<15}"

        size = ','.join(split[:-5])
        channels_last, mode, antialias, dtype, threads= split[-5:]
        deets = f"{size:<33} {channels_last:<7} {antialias:<7} {mode:<10} {threads:<15}"

        l = f"{deets}  {improv:<5} {main_time / 1000:{time_fmt}}ms vs {new_time / 1000:{time_fmt}}ms"
        out.append(l)


def key(s):
    # s = ''.join(s.split()[1:]) # remove "N.nX" part
    num_threads = (int(re.findall(r"num_threads=(\d+)", s)[0]),)

    input_shape, output_shape = re.findall("\(.*?\)", s)
    input_shape = input_shape[1:-1]  # remove parenthesis
    input_HW = tuple(int(x) for x in input_shape.split(",")[-2:])
    input_C = (-int(input_shape.split(",")[1]),)

    output_HW = tuple(int(x) for x in output_shape[1:-1].split(","))
    is_downsample = (output_HW[0] < input_HW[0],)
    if "linear" in s:
        mode = "linear"
    elif "nearest-exact" in s:
        mode = "nearest-exact"
    else:
        # assert "nearest" in s
        mode = "nearest"
    mode = (mode,)
    return is_downsample + input_HW + output_HW + num_threads + input_C + mode

for i, l in enumerate(sorted(out, key=key)):
    if i % 8 == 0:
        print()
    # if i % 10 == 0 and i % 40 != 0:
    #     print()
    # if i % 40 == 0:
    #     print("-" * 100)
    print(l)

cc @jgong5 @mingfeima @XiaobingSuper @sanchitintel @ashokei @jingxu10 @datumbox @vfdev-5 @pmeier

…t_dev

…6_weights_aa

…as_false' into interpolate_uint8_images_linear_cpu_support_dev

pytorch-bot · 2022-12-13T16:13:37Z

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/90771

📄 Preview Python docs built from this PR
📄 Preview C++ docs built from this PR
❓ Need help or want to give feedback on the CI? Visit the bot commands wiki or our office hours

Note: Links to docs will display an error until the docs builds have been completed.

❗ 1 Active SEVs

There are 1 currently active SEVs. If your PR is affected, please view them below:

ROCm jobs fail to access AMD apt repo

✅ No Failures

As of commit 35443d0:
💚 Looks good so far! There are no failures yet. 💚

This comment was automatically generated by Dr. CI and updates every 15 minutes.

vadimkantorov · 2022-12-13T16:33:09Z

Related: #5580
Somewhat related: #54389 (lerp for uint8 - so interpolation, but in another sense :))

vadimkantorov · 2022-12-13T16:36:14Z

Maybe support for bool+nearest can also be enabled without much hassle (especially given that uint8+nearest is already supported). I also wonder, for bool specifically, are all other modes equivalent to nearest?

Bool interpolation is useful when working with segmentation masks

vadimkantorov · 2022-12-13T19:36:45Z

@NicolasHug There're a bunch of perf benchmarks of basic transforms in Albumentations repo: https://github.com/albumentations-team/albumentations#benchmarking-results

It would be interesting to have such benchmarks run by pytorch itself (to ensure the correct version of torchvision is used etc) and having the results published. Maybe also contribute to the original albumentations benchmarking code.

https://github.com/albumentations-team/albumentations/blob/master/benchmark/README.md

It would be interesting to see the basic transforms for which core torch / torchvision is slower than albumentations

vfdev-5 · 2022-12-13T20:48:25Z

@vadimkantorov albu uses opencv for majority of transforms and opencv is highly optimized. On the other hand if we can achieve Pillow-SIMD runtime results in pytorch it would be good.
While benchmarking, we also need to take into account that we measure the same ops, for example resize downsampling in opencv does not use antialiasing and Pillow uses antialiasing.

## Description - Based on #96651 - Improved perfs for vectorized interpolate uint8 RGB-case - unified RGB and RGBA processing code such that RGB input is not copied into RGBA - Performances are more close to Pillow-SIMD - RGBA case perfs are the same after refactoring (see Source link below) - Fixed mem pointer alignment, added more comments (reviews from #96651) ## Results ``` [------------------------------------------------------------------------------------------ Resize -----------------------------------------------------------------------------------------] | Pillow (9.0.0.post1) | torch (2.1.0a0+gitcc42a3f) PR | torch (2.1.0a0+git5309c44) nightly | Speed-up: PR vs nightly 1 threads: ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 3 torch.uint8 channels_last bilinear 256 -> 32 aa=True | 38.8 | 56.0 | 133.2 | 2.4 3 torch.uint8 channels_last bilinear 256 -> 32 aa=False | | 37.5 | 112.8 | 3.0 3 torch.uint8 channels_last bilinear 256 -> 224 aa=True | 128.7 | 157.0 | 305.4 | 1.9 3 torch.uint8 channels_last bilinear 256 -> 224 aa=False | | 146.4 | 288.7 | 2.0 3 torch.uint8 channels_last bilinear 256 -> 320 aa=True | 179.4 | 215.8 | 442.5 | 2.1 3 torch.uint8 channels_last bilinear 256 -> 320 aa=False | | 212.5 | 436.9 | 2.1 3 torch.uint8 channels_last bilinear 520 -> 32 aa=True | 113.3 | 127.9 | 464.8 | 3.6 3 torch.uint8 channels_last bilinear 520 -> 32 aa=False | | 56.8 | 365.5 | 6.4 3 torch.uint8 channels_last bilinear 520 -> 224 aa=True | 281.7 | 325.2 | 722.4 | 2.2 3 torch.uint8 channels_last bilinear 520 -> 224 aa=False | | 239.1 | 593.5 | 2.5 3 torch.uint8 channels_last bilinear 712 -> 32 aa=True | 186.2 | 200.7 | 833.8 | 4.2 3 torch.uint8 channels_last bilinear 712 -> 32 aa=False | | 75.2 | 651.4 | 8.7 3 torch.uint8 channels_last bilinear 712 -> 224 aa=True | 410.0 | 444.5 | 1128.4 | 2.5 3 torch.uint8 channels_last bilinear 712 -> 224 aa=False | | 309.3 | 917.6 | 3.0 ``` Note: for other cases (see Source below) speed-up is roughly around 1.0 +/- 0.1 which may be attributed to noisy measurements ... [Source](https://gist.github.com/vfdev-5/1c0778904a07ce40401306548b9525e8#file-20230315-144416-pr_vs_nightly_speedup-md) ## Context - #90771 cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 [ghstack-poisoned]

…cpu uint8 RGB-case" ## Description - Based on #96651 - Improved perfs for vectorized interpolate uint8 RGB-case - unified RGB and RGBA processing code such that RGB input is not copied into RGBA - Performances are more close to Pillow-SIMD - RGBA case perfs are the same after refactoring (see Source link below) - Fixed mem pointer alignment, added more comments (reviews from #96651) ## Results ``` [------------------------------------------------------------------------------------------ Resize -----------------------------------------------------------------------------------------] | Pillow (9.0.0.post1) | torch (2.1.0a0+git0968a5d) PR | torch (2.1.0a0+git5309c44) nightly | Speed-up: PR vs nightly 1 threads: ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 3 torch.uint8 channels_last bilinear 256 -> 32 aa=True | 39.0 | 56.6 | 133.2 | 2.4 3 torch.uint8 channels_last bilinear 256 -> 32 aa=False | | 36.9 | 112.8 | 3.1 3 torch.uint8 channels_last bilinear 256 -> 224 aa=True | 128.1 | 152.5 | 305.4 | 2.0 3 torch.uint8 channels_last bilinear 256 -> 224 aa=False | | 141.1 | 288.7 | 2.0 3 torch.uint8 channels_last bilinear 256 -> 320 aa=True | 179.6 | 208.8 | 442.5 | 2.1 3 torch.uint8 channels_last bilinear 256 -> 320 aa=False | | 206.4 | 436.9 | 2.1 3 torch.uint8 channels_last bilinear 520 -> 32 aa=True | 113.3 | 132.1 | 464.8 | 3.5 3 torch.uint8 channels_last bilinear 520 -> 32 aa=False | | 57.2 | 365.5 | 6.4 3 torch.uint8 channels_last bilinear 520 -> 224 aa=True | 281.7 | 327.4 | 722.4 | 2.2 3 torch.uint8 channels_last bilinear 520 -> 224 aa=False | | 230.2 | 593.5 | 2.6 3 torch.uint8 channels_last bilinear 712 -> 32 aa=True | 186.9 | 210.5 | 833.8 | 4.0 3 torch.uint8 channels_last bilinear 712 -> 32 aa=False | | 75.6 | 651.4 | 8.6 3 torch.uint8 channels_last bilinear 712 -> 224 aa=True | 410.3 | 450.9 | 1128.4 | 2.5 3 torch.uint8 channels_last bilinear 712 -> 224 aa=False | | 298.7 | 917.6 | 3.1 ``` Note: for other cases (see Source below) speed-up is roughly around 1.0 +/- 0.1 which may be attributed to noisy measurements ... [Source](https://gist.github.com/vfdev-5/1c0778904a07ce40401306548b9525e8#file-20230315-162238-pr_vs_nightly_speedup-md) ## Context - #90771 cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 [ghstack-poisoned]

## Description - Based on #96651 - Improved perfs for vectorized interpolate uint8 RGB-case - unified RGB and RGBA processing code such that RGB input is not copied into RGBA - Performances are more close to Pillow-SIMD - RGBA case perfs are the same after refactoring (see Source link below) - Fixed mem pointer alignment, added more comments (reviews from #96651) ## Results ``` [------------------------------------------------------------------------------------------ Resize -----------------------------------------------------------------------------------------] | Pillow (9.0.0.post1) | torch (2.1.0a0+git0968a5d) PR | torch (2.1.0a0+git5309c44) nightly | Speed-up: PR vs nightly 1 threads: ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 3 torch.uint8 channels_last bilinear 256 -> 32 aa=True | 39.0 | 56.6 | 133.2 | 2.4 3 torch.uint8 channels_last bilinear 256 -> 32 aa=False | | 36.9 | 112.8 | 3.1 3 torch.uint8 channels_last bilinear 256 -> 224 aa=True | 128.1 | 152.5 | 305.4 | 2.0 3 torch.uint8 channels_last bilinear 256 -> 224 aa=False | | 141.1 | 288.7 | 2.0 3 torch.uint8 channels_last bilinear 256 -> 320 aa=True | 179.6 | 208.8 | 442.5 | 2.1 3 torch.uint8 channels_last bilinear 256 -> 320 aa=False | | 206.4 | 436.9 | 2.1 3 torch.uint8 channels_last bilinear 520 -> 32 aa=True | 113.3 | 132.1 | 464.8 | 3.5 3 torch.uint8 channels_last bilinear 520 -> 32 aa=False | | 57.2 | 365.5 | 6.4 3 torch.uint8 channels_last bilinear 520 -> 224 aa=True | 281.7 | 327.4 | 722.4 | 2.2 3 torch.uint8 channels_last bilinear 520 -> 224 aa=False | | 230.2 | 593.5 | 2.6 3 torch.uint8 channels_last bilinear 712 -> 32 aa=True | 186.9 | 210.5 | 833.8 | 4.0 3 torch.uint8 channels_last bilinear 712 -> 32 aa=False | | 75.6 | 651.4 | 8.6 3 torch.uint8 channels_last bilinear 712 -> 224 aa=True | 410.3 | 450.9 | 1128.4 | 2.5 3 torch.uint8 channels_last bilinear 712 -> 224 aa=False | | 298.7 | 917.6 | 3.1 ``` Note: for other cases (see Source below) speed-up is roughly around 1.0 +/- 0.1 which may be attributed to noisy measurements ... [Source](https://gist.github.com/vfdev-5/1c0778904a07ce40401306548b9525e8#file-20230315-162238-pr_vs_nightly_speedup-md) ## Context - #90771 cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 [ghstack-poisoned]

…erpolate cpu uint8 RGB-case (channels last)" ## Description - Based on #96651 - Improved perfs for vectorized bilinear interpolate uint8 RGB-case, channels last - unified RGB and RGBA processing code such that RGB input is not copied into RGBA - Performances are more close to Pillow-SIMD (`Pillow (9.0.0.post1)`) - RGBA case perfs are the same after refactoring (see Source link below) - Fixed mem pointer alignment, added more comments (reviews from #96651) ## Results ``` [------------------------------------------------------------------------------------------ Resize -----------------------------------------------------------------------------------------] | Pillow (9.0.0.post1) | torch (2.1.0a0+git0968a5d) PR | torch (2.1.0a0+git5309c44) nightly | Speed-up: PR vs nightly 1 threads: ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 3 torch.uint8 channels_last bilinear 256 -> 32 aa=True | 39.0 | 56.6 | 133.2 | 2.4 3 torch.uint8 channels_last bilinear 256 -> 32 aa=False | | 36.9 | 112.8 | 3.1 3 torch.uint8 channels_last bilinear 256 -> 224 aa=True | 128.1 | 152.5 | 305.4 | 2.0 3 torch.uint8 channels_last bilinear 256 -> 224 aa=False | | 141.1 | 288.7 | 2.0 3 torch.uint8 channels_last bilinear 256 -> 320 aa=True | 179.6 | 208.8 | 442.5 | 2.1 3 torch.uint8 channels_last bilinear 256 -> 320 aa=False | | 206.4 | 436.9 | 2.1 3 torch.uint8 channels_last bilinear 520 -> 32 aa=True | 113.3 | 132.1 | 464.8 | 3.5 3 torch.uint8 channels_last bilinear 520 -> 32 aa=False | | 57.2 | 365.5 | 6.4 3 torch.uint8 channels_last bilinear 520 -> 224 aa=True | 281.7 | 327.4 | 722.4 | 2.2 3 torch.uint8 channels_last bilinear 520 -> 224 aa=False | | 230.2 | 593.5 | 2.6 3 torch.uint8 channels_last bilinear 712 -> 32 aa=True | 186.9 | 210.5 | 833.8 | 4.0 3 torch.uint8 channels_last bilinear 712 -> 32 aa=False | | 75.6 | 651.4 | 8.6 3 torch.uint8 channels_last bilinear 712 -> 224 aa=True | 410.3 | 450.9 | 1128.4 | 2.5 3 torch.uint8 channels_last bilinear 712 -> 224 aa=False | | 298.7 | 917.6 | 3.1 ``` Note: for other cases (see Source below) speed-up is roughly around 1.0 +/- 0.1 which may be attributed to noisy measurements ... [Source](https://gist.github.com/vfdev-5/1c0778904a07ce40401306548b9525e8#file-20230315-162238-pr_vs_nightly_speedup-md) ## Context - #90771 cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 [ghstack-poisoned]

…t8 RGB-case (channels last)" ## Description - Based on #96651 - Improved perfs for vectorized bilinear interpolate uint8 RGB-case, channels last - unified RGB and RGBA processing code such that RGB input is not copied into RGBA - Performances are more close to Pillow-SIMD (`Pillow (9.0.0.post1)`) - RGBA case perfs are the same after refactoring (see Source link below) - Fixed mem pointer alignment, added more comments (reviews from #96651) ## Results ``` [------------------------------------------------------------------------------------------ Resize -----------------------------------------------------------------------------------------] | Pillow (9.0.0.post1) | torch (2.1.0a0+git0968a5d) PR | torch (2.1.0a0+git5309c44) nightly | Speed-up: PR vs nightly 1 threads: ---------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 3 torch.uint8 channels_last bilinear 256 -> 32 aa=True | 39.0 | 56.6 | 133.2 | 2.4 3 torch.uint8 channels_last bilinear 256 -> 32 aa=False | | 36.9 | 112.8 | 3.1 3 torch.uint8 channels_last bilinear 256 -> 224 aa=True | 128.1 | 152.5 | 305.4 | 2.0 3 torch.uint8 channels_last bilinear 256 -> 224 aa=False | | 141.1 | 288.7 | 2.0 3 torch.uint8 channels_last bilinear 256 -> 320 aa=True | 179.6 | 208.8 | 442.5 | 2.1 3 torch.uint8 channels_last bilinear 256 -> 320 aa=False | | 206.4 | 436.9 | 2.1 3 torch.uint8 channels_last bilinear 520 -> 32 aa=True | 113.3 | 132.1 | 464.8 | 3.5 3 torch.uint8 channels_last bilinear 520 -> 32 aa=False | | 57.2 | 365.5 | 6.4 3 torch.uint8 channels_last bilinear 520 -> 224 aa=True | 281.7 | 327.4 | 722.4 | 2.2 3 torch.uint8 channels_last bilinear 520 -> 224 aa=False | | 230.2 | 593.5 | 2.6 3 torch.uint8 channels_last bilinear 712 -> 32 aa=True | 186.9 | 210.5 | 833.8 | 4.0 3 torch.uint8 channels_last bilinear 712 -> 32 aa=False | | 75.6 | 651.4 | 8.6 3 torch.uint8 channels_last bilinear 712 -> 224 aa=True | 410.3 | 450.9 | 1128.4 | 2.5 3 torch.uint8 channels_last bilinear 712 -> 224 aa=False | | 298.7 | 917.6 | 3.1 ``` Note: for other cases (see Source below) speed-up is roughly around 1.0 +/- 0.1 which may be attributed to noisy measurements ... [Source](https://gist.github.com/vfdev-5/1c0778904a07ce40401306548b9525e8#file-20230315-162238-pr_vs_nightly_speedup-md) ## Context - #90771 cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 [ghstack-poisoned]

…erpolate cpu uint8 RGB-case (channels last)" ## Description - Based on #96651 - Improved perfs for vectorized **bilinear** interpolate uint8 RGB-case, **channels last** - unified RGB and RGBA processing code such that RGB input is not copied into RGBA - Performances are more close to Pillow-SIMD (labeled as `Pillow (9.0.0.post1)` in the results) - RGBA case perfs are the same after refactoring (see Source link below) - Fixed mem pointer alignment, added more comments (reviews from #96651) ## Results - `Pillow (9.0.0.post1)` == Pillow-SIMD ``` [-------------------------------------------------------------------------------------------------- Resize -------------------------------------------------------------------------------------------------] | Pillow (9.0.0.post1) | torch (2.1.0a0+gitc005105) PR | torch (2.1.0a0+git5309c44) nightly | Speed-up: PR vs nightly 1 threads: -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=True | 38.670 (+-0.445) | 57.366 (+-0.799) | 132.147 (+-1.236) | 2.304 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=False | | 37.825 (+-0.417) | 111.789 (+-1.175) | 2.955 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=True | 127.898 (+-1.335) | 153.081 (+-2.346) | 302.518 (+-2.632) | 1.976 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=False | | 141.695 (+-1.415) | 286.663 (+-2.494) | 2.023 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=True | 179.735 (+-2.054) | 210.613 (+-3.116) | 439.375 (+-4.014) | 2.086 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=False | | 207.601 (+-1.639) | 438.537 (+-4.143) | 2.112 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=True | 112.679 (+-1.321) | 130.863 (+-1.987) | 446.804 (+-3.283) | 3.414 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=False | | 57.968 (+-0.270) | 374.244 (+-13.598) | 6.456 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=True | 282.398 (+-3.485) | 322.986 (+-1.947) | 720.197 (+-3.467) | 2.230 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=False | | 231.625 (+-2.006) | 592.834 (+-3.903) | 2.559 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=True | 185.711 (+-1.666) | 201.069 (+-2.182) | 787.868 (+-3.648) | 3.918 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=False | | 75.975 (+-0.696) | 651.016 (+-3.926) | 8.569 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=True | 410.236 (+-6.021) | 451.486 (+-3.939) | 1123.923 (+-14.988) | 2.489 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=False | | 299.597 (+-1.887) | 915.347 (+-4.486) | 3.055 (+-0.000) # More test-cases from #90771 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=True | 60.751 (+-0.285) | 78.538 (+-1.282) | 170.465 (+-1.830) | 2.170 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=True | 133.619 (+-2.035) | 159.614 (+-1.587) | 330.971 (+-3.249) | 2.074 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=True | 950.243 (+-10.641) | 891.369 (+-17.946) | 2805.510 (+-25.503) | 3.147 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=True | 52.771 (+-0.961) | 72.253 (+-1.020) | 135.933 (+-1.625) | 1.881 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=True | 139.107 (+-2.143) | 165.844 (+-2.177) | 321.112 (+-2.904) | 1.936 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=True | 691.470 (+-9.566) | 764.942 (+-11.192) | 2050.880 (+-22.188) | 2.681 (+-0.000) 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=False | | 77.375 (+-1.345) | 169.646 (+-1.640) | 2.193 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=False | | 159.115 (+-3.935) | 329.754 (+-2.590) | 2.072 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=False | | 877.248 (+-5.736) | 2815.870 (+-22.589) | 3.210 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=False | | 53.120 (+-0.316) | 112.024 (+-1.225) | 2.109 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=False | | 147.330 (+-1.871) | 299.152 (+-3.353) | 2.030 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=False | | 472.182 (+-10.785) | 1698.601 (+-16.785) | 3.597 (+-0.000) ``` Note: for other cases (see Source below) speed-up is roughly around 1.0 +/- 0.1 which may be attributed to noisy measurements ... [Source](https://gist.github.com/vfdev-5/1c0778904a07ce40401306548b9525e8#file-20230320-160044-pr_vs_nightly-speedup-md) ## Context - #90771 cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 [ghstack-poisoned]

…t8 RGB-case (channels last)" ## Description - Based on #96651 - Improved perfs for vectorized **bilinear** interpolate uint8 RGB-case, **channels last** - unified RGB and RGBA processing code such that RGB input is not copied into RGBA - Performances are more close to Pillow-SIMD (labeled as `Pillow (9.0.0.post1)` in the results) - RGBA case perfs are the same after refactoring (see Source link below) - Fixed mem pointer alignment, added more comments (reviews from #96651) ## Results - `Pillow (9.0.0.post1)` == Pillow-SIMD ``` [-------------------------------------------------------------------------------------------------- Resize -------------------------------------------------------------------------------------------------] | Pillow (9.0.0.post1) | torch (2.1.0a0+gitc005105) PR | torch (2.1.0a0+git5309c44) nightly | Speed-up: PR vs nightly 1 threads: -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=True | 38.670 (+-0.445) | 57.366 (+-0.799) | 132.147 (+-1.236) | 2.304 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=False | | 37.825 (+-0.417) | 111.789 (+-1.175) | 2.955 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=True | 127.898 (+-1.335) | 153.081 (+-2.346) | 302.518 (+-2.632) | 1.976 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=False | | 141.695 (+-1.415) | 286.663 (+-2.494) | 2.023 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=True | 179.735 (+-2.054) | 210.613 (+-3.116) | 439.375 (+-4.014) | 2.086 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=False | | 207.601 (+-1.639) | 438.537 (+-4.143) | 2.112 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=True | 112.679 (+-1.321) | 130.863 (+-1.987) | 446.804 (+-3.283) | 3.414 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=False | | 57.968 (+-0.270) | 374.244 (+-13.598) | 6.456 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=True | 282.398 (+-3.485) | 322.986 (+-1.947) | 720.197 (+-3.467) | 2.230 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=False | | 231.625 (+-2.006) | 592.834 (+-3.903) | 2.559 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=True | 185.711 (+-1.666) | 201.069 (+-2.182) | 787.868 (+-3.648) | 3.918 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=False | | 75.975 (+-0.696) | 651.016 (+-3.926) | 8.569 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=True | 410.236 (+-6.021) | 451.486 (+-3.939) | 1123.923 (+-14.988) | 2.489 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=False | | 299.597 (+-1.887) | 915.347 (+-4.486) | 3.055 (+-0.000) # More test-cases from #90771 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=True | 60.751 (+-0.285) | 78.538 (+-1.282) | 170.465 (+-1.830) | 2.170 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=True | 133.619 (+-2.035) | 159.614 (+-1.587) | 330.971 (+-3.249) | 2.074 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=True | 950.243 (+-10.641) | 891.369 (+-17.946) | 2805.510 (+-25.503) | 3.147 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=True | 52.771 (+-0.961) | 72.253 (+-1.020) | 135.933 (+-1.625) | 1.881 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=True | 139.107 (+-2.143) | 165.844 (+-2.177) | 321.112 (+-2.904) | 1.936 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=True | 691.470 (+-9.566) | 764.942 (+-11.192) | 2050.880 (+-22.188) | 2.681 (+-0.000) 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=False | | 77.375 (+-1.345) | 169.646 (+-1.640) | 2.193 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=False | | 159.115 (+-3.935) | 329.754 (+-2.590) | 2.072 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=False | | 877.248 (+-5.736) | 2815.870 (+-22.589) | 3.210 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=False | | 53.120 (+-0.316) | 112.024 (+-1.225) | 2.109 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=False | | 147.330 (+-1.871) | 299.152 (+-3.353) | 2.030 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=False | | 472.182 (+-10.785) | 1698.601 (+-16.785) | 3.597 (+-0.000) ``` Note: for other cases (see Source below) speed-up is roughly around 1.0 +/- 0.1 which may be attributed to noisy measurements ... [Source](https://gist.github.com/vfdev-5/1c0778904a07ce40401306548b9525e8#file-20230320-160044-pr_vs_nightly-speedup-md) ## Context - #90771 cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 [ghstack-poisoned]

…erpolate cpu uint8 RGB-case (channels last)" ## Description - Based on #96651 - Improved perfs for vectorized **bilinear** interpolate uint8 RGB-case, **channels last** - unified RGB and RGBA processing code such that RGB input is not copied into RGBA - Performances are more close to Pillow-SIMD (labeled as `Pillow (9.0.0.post1)` in the results) - RGBA case perfs are the same after refactoring (see Source link below) - Fixed mem pointer alignment, added more comments (reviews from #96651) ## Results - `Pillow (9.0.0.post1)` == Pillow-SIMD ``` [-------------------------------------------------------------------------------------------------- Resize -------------------------------------------------------------------------------------------------] | Pillow (9.0.0.post1) | torch (2.1.0a0+git8d955df) PR | torch (2.1.0a0+git5309c44) nightly | Speed-up: PR vs nightly 1 threads: -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=True | 38.649 (+-0.306) | 55.828 (+-0.370) | 132.147 (+-1.236) | 2.367 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=False | | 36.826 (+-0.229) | 111.789 (+-1.175) | 3.036 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=True | 128.233 (+-1.313) | 153.827 (+-1.229) | 302.518 (+-2.632) | 1.967 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=False | | 143.886 (+-1.409) | 286.663 (+-2.494) | 1.992 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=True | 179.504 (+-1.825) | 211.569 (+-1.336) | 439.375 (+-4.014) | 2.077 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=False | | 209.888 (+-1.443) | 438.537 (+-4.143) | 2.089 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=True | 112.891 (+-1.118) | 129.373 (+-1.396) | 446.804 (+-3.283) | 3.454 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=False | | 56.858 (+-0.227) | 374.244 (+-13.598) | 6.582 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=True | 282.917 (+-2.992) | 324.378 (+-1.694) | 720.197 (+-3.467) | 2.220 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=False | | 236.078 (+-1.679) | 592.834 (+-3.903) | 2.511 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=True | 185.595 (+-1.633) | 202.000 (+-1.920) | 787.868 (+-3.648) | 3.900 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=False | | 75.421 (+-0.512) | 651.016 (+-3.926) | 8.632 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=True | 409.691 (+-2.735) | 449.927 (+-2.500) | 1123.923 (+-14.988) | 2.498 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=False | | 306.691 (+-2.095) | 915.347 (+-4.486) | 2.985 (+-0.000) # More test-cases from #90771 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=True | 60.740 (+-0.278) | 78.745 (+-0.286) | 170.465 (+-1.830) | 2.165 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=True | 133.029 (+-1.619) | 162.393 (+-1.289) | 330.971 (+-3.249) | 2.038 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=True | 948.849 (+-2.749) | 896.127 (+-3.696) | 2805.510 (+-25.503) | 3.131 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=True | 52.505 (+-0.319) | 70.617 (+-0.344) | 135.933 (+-1.625) | 1.925 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=True | 138.671 (+-1.953) | 165.638 (+-1.473) | 321.112 (+-2.904) | 1.939 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=True | 689.492 (+-2.917) | 758.162 (+-3.719) | 2050.880 (+-22.188) | 2.705 (+-0.000) 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=False | | 77.300 (+-0.307) | 169.646 (+-1.640) | 2.195 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=False | | 159.525 (+-1.225) | 329.754 (+-2.590) | 2.067 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=False | | 890.106 (+-3.358) | 2815.870 (+-22.589) | 3.164 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=False | | 52.399 (+-0.314) | 112.024 (+-1.225) | 2.138 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=False | | 148.780 (+-1.282) | 299.152 (+-3.353) | 2.011 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=False | | 479.273 (+-3.432) | 1698.601 (+-16.785) | 3.544 (+-0.000) 4 ``` Note: There is no perf regression for other case. There some cases (see Source below) with small speed-ups, for the rest it is roughly around 1.0 +/- 0.1 which may be attributed to noisy measurements ... [Source](https://gist.github.com/vfdev-5/1c0778904a07ce40401306548b9525e8#file-20230321-145513-pr_vs_nightly-speedup-md) ## Context - #90771 cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 [ghstack-poisoned]

…t8 RGB-case (channels last)" ## Description - Based on #96651 - Improved perfs for vectorized **bilinear** interpolate uint8 RGB-case, **channels last** - unified RGB and RGBA processing code such that RGB input is not copied into RGBA - Performances are more close to Pillow-SIMD (labeled as `Pillow (9.0.0.post1)` in the results) - RGBA case perfs are the same after refactoring (see Source link below) - Fixed mem pointer alignment, added more comments (reviews from #96651) ## Results - `Pillow (9.0.0.post1)` == Pillow-SIMD ``` [-------------------------------------------------------------------------------------------------- Resize -------------------------------------------------------------------------------------------------] | Pillow (9.0.0.post1) | torch (2.1.0a0+git8d955df) PR | torch (2.1.0a0+git5309c44) nightly | Speed-up: PR vs nightly 1 threads: -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=True | 38.649 (+-0.306) | 55.828 (+-0.370) | 132.147 (+-1.236) | 2.367 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=False | | 36.826 (+-0.229) | 111.789 (+-1.175) | 3.036 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=True | 128.233 (+-1.313) | 153.827 (+-1.229) | 302.518 (+-2.632) | 1.967 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=False | | 143.886 (+-1.409) | 286.663 (+-2.494) | 1.992 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=True | 179.504 (+-1.825) | 211.569 (+-1.336) | 439.375 (+-4.014) | 2.077 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=False | | 209.888 (+-1.443) | 438.537 (+-4.143) | 2.089 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=True | 112.891 (+-1.118) | 129.373 (+-1.396) | 446.804 (+-3.283) | 3.454 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=False | | 56.858 (+-0.227) | 374.244 (+-13.598) | 6.582 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=True | 282.917 (+-2.992) | 324.378 (+-1.694) | 720.197 (+-3.467) | 2.220 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=False | | 236.078 (+-1.679) | 592.834 (+-3.903) | 2.511 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=True | 185.595 (+-1.633) | 202.000 (+-1.920) | 787.868 (+-3.648) | 3.900 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=False | | 75.421 (+-0.512) | 651.016 (+-3.926) | 8.632 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=True | 409.691 (+-2.735) | 449.927 (+-2.500) | 1123.923 (+-14.988) | 2.498 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=False | | 306.691 (+-2.095) | 915.347 (+-4.486) | 2.985 (+-0.000) # More test-cases from #90771 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=True | 60.740 (+-0.278) | 78.745 (+-0.286) | 170.465 (+-1.830) | 2.165 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=True | 133.029 (+-1.619) | 162.393 (+-1.289) | 330.971 (+-3.249) | 2.038 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=True | 948.849 (+-2.749) | 896.127 (+-3.696) | 2805.510 (+-25.503) | 3.131 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=True | 52.505 (+-0.319) | 70.617 (+-0.344) | 135.933 (+-1.625) | 1.925 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=True | 138.671 (+-1.953) | 165.638 (+-1.473) | 321.112 (+-2.904) | 1.939 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=True | 689.492 (+-2.917) | 758.162 (+-3.719) | 2050.880 (+-22.188) | 2.705 (+-0.000) 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=False | | 77.300 (+-0.307) | 169.646 (+-1.640) | 2.195 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=False | | 159.525 (+-1.225) | 329.754 (+-2.590) | 2.067 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=False | | 890.106 (+-3.358) | 2815.870 (+-22.589) | 3.164 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=False | | 52.399 (+-0.314) | 112.024 (+-1.225) | 2.138 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=False | | 148.780 (+-1.282) | 299.152 (+-3.353) | 2.011 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=False | | 479.273 (+-3.432) | 1698.601 (+-16.785) | 3.544 (+-0.000) 4 ``` Note: There is no perf regression for other case. There some cases (see Source below) with small speed-ups, for the rest it is roughly around 1.0 +/- 0.1 which may be attributed to noisy measurements ... [Source](https://gist.github.com/vfdev-5/1c0778904a07ce40401306548b9525e8#file-20230321-145513-pr_vs_nightly-speedup-md) ## Context - #90771 cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 [ghstack-poisoned]

…erpolate cpu uint8 RGB-case (channels last)" ## Description - Based on #96651 - Improved perfs for vectorized **bilinear** interpolate uint8 RGB-case, **channels last** - unified RGB and RGBA processing code such that RGB input is not copied into RGBA - Performances are more close to Pillow-SIMD (labeled as `Pillow (9.0.0.post1)` in the results) - RGBA case perfs are the same after refactoring (see Source link below) - Fixed mem pointer alignment, added more comments (reviews from #96651) ## Results - `Pillow (9.0.0.post1)` == Pillow-SIMD ``` [-------------------------------------------------------------------------------------------------- Resize -------------------------------------------------------------------------------------------------] | Pillow (9.0.0.post1) | torch (2.1.0a0+gitce4be01) PR | torch (2.1.0a0+git5309c44) nightly | Speed-up: PR vs nightly 1 threads: -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=True | 38.548 (+-0.280) | 57.536 (+-0.210) | 132.147 (+-1.236) | 2.297 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=False | | 38.532 (+-0.219) | 111.789 (+-1.175) | 2.901 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=True | 127.689 (+-1.348) | 156.262 (+-1.213) | 302.518 (+-2.632) | 1.936 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=False | | 145.483 (+-1.077) | 286.663 (+-2.494) | 1.970 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=True | 178.117 (+-1.956) | 215.053 (+-1.470) | 439.375 (+-4.014) | 2.043 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=False | | 211.340 (+-2.239) | 438.537 (+-4.143) | 2.075 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=True | 112.593 (+-1.266) | 130.414 (+-1.633) | 446.804 (+-3.283) | 3.426 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=False | | 58.767 (+-0.203) | 374.244 (+-13.598) | 6.368 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=True | 283.210 (+-2.937) | 324.157 (+-1.895) | 720.197 (+-3.467) | 2.222 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=False | | 239.800 (+-2.492) | 592.834 (+-3.903) | 2.472 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=True | 186.255 (+-1.629) | 204.834 (+-1.496) | 787.868 (+-3.648) | 3.846 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=False | | 77.335 (+-0.341) | 651.016 (+-3.926) | 8.418 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=True | 410.286 (+-2.439) | 443.934 (+-2.899) | 1123.923 (+-14.988) | 2.532 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=False | | 312.220 (+-2.307) | 915.347 (+-4.486) | 2.932 (+-0.000) # More test-cases from #90771 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=True | 60.611 (+-0.337) | 80.849 (+-1.780) | 170.465 (+-1.830) | 2.108 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=True | 132.971 (+-1.624) | 164.892 (+-1.426) | 330.971 (+-3.249) | 2.007 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=True | 948.467 (+-3.179) | 891.414 (+-5.282) | 2805.510 (+-25.503) | 3.147 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=True | 52.539 (+-0.327) | 72.471 (+-0.367) | 135.933 (+-1.625) | 1.876 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=True | 138.669 (+-1.867) | 168.628 (+-1.213) | 321.112 (+-2.904) | 1.904 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=True | 689.933 (+-3.175) | 746.911 (+-2.985) | 2050.880 (+-22.188) | 2.746 (+-0.000) 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=False | | 78.347 (+-0.338) | 169.646 (+-1.640) | 2.165 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=False | | 162.194 (+-1.089) | 329.754 (+-2.590) | 2.033 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=False | | 894.476 (+-2.738) | 2815.870 (+-22.589) | 3.148 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=False | | 52.728 (+-0.406) | 112.024 (+-1.225) | 2.125 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=False | | 151.560 (+-1.128) | 299.152 (+-3.353) | 1.974 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=False | | 500.053 (+-4.288) | 1698.601 (+-16.785) | 3.397 (+-0.000) ``` Note: There is no perf regression for other case. There some cases (see Source below) with small speed-ups, for the rest it is roughly around 1.0 +/- 0.1 which may be attributed to noisy measurements ... [Source](https://gist.github.com/vfdev-5/1c0778904a07ce40401306548b9525e8#file-20230322-132441-pr_vs_nightly-speedup-md) ## Context - #90771 cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 [ghstack-poisoned]

…t8 RGB-case (channels last)" ## Description - Based on #96651 - Improved perfs for vectorized **bilinear** interpolate uint8 RGB-case, **channels last** - unified RGB and RGBA processing code such that RGB input is not copied into RGBA - Performances are more close to Pillow-SIMD (labeled as `Pillow (9.0.0.post1)` in the results) - RGBA case perfs are the same after refactoring (see Source link below) - Fixed mem pointer alignment, added more comments (reviews from #96651) ## Results - `Pillow (9.0.0.post1)` == Pillow-SIMD ``` [-------------------------------------------------------------------------------------------------- Resize -------------------------------------------------------------------------------------------------] | Pillow (9.0.0.post1) | torch (2.1.0a0+gitce4be01) PR | torch (2.1.0a0+git5309c44) nightly | Speed-up: PR vs nightly 1 threads: -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=True | 38.548 (+-0.280) | 57.536 (+-0.210) | 132.147 (+-1.236) | 2.297 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=False | | 38.532 (+-0.219) | 111.789 (+-1.175) | 2.901 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=True | 127.689 (+-1.348) | 156.262 (+-1.213) | 302.518 (+-2.632) | 1.936 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=False | | 145.483 (+-1.077) | 286.663 (+-2.494) | 1.970 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=True | 178.117 (+-1.956) | 215.053 (+-1.470) | 439.375 (+-4.014) | 2.043 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=False | | 211.340 (+-2.239) | 438.537 (+-4.143) | 2.075 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=True | 112.593 (+-1.266) | 130.414 (+-1.633) | 446.804 (+-3.283) | 3.426 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=False | | 58.767 (+-0.203) | 374.244 (+-13.598) | 6.368 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=True | 283.210 (+-2.937) | 324.157 (+-1.895) | 720.197 (+-3.467) | 2.222 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=False | | 239.800 (+-2.492) | 592.834 (+-3.903) | 2.472 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=True | 186.255 (+-1.629) | 204.834 (+-1.496) | 787.868 (+-3.648) | 3.846 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=False | | 77.335 (+-0.341) | 651.016 (+-3.926) | 8.418 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=True | 410.286 (+-2.439) | 443.934 (+-2.899) | 1123.923 (+-14.988) | 2.532 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=False | | 312.220 (+-2.307) | 915.347 (+-4.486) | 2.932 (+-0.000) # More test-cases from #90771 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=True | 60.611 (+-0.337) | 80.849 (+-1.780) | 170.465 (+-1.830) | 2.108 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=True | 132.971 (+-1.624) | 164.892 (+-1.426) | 330.971 (+-3.249) | 2.007 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=True | 948.467 (+-3.179) | 891.414 (+-5.282) | 2805.510 (+-25.503) | 3.147 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=True | 52.539 (+-0.327) | 72.471 (+-0.367) | 135.933 (+-1.625) | 1.876 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=True | 138.669 (+-1.867) | 168.628 (+-1.213) | 321.112 (+-2.904) | 1.904 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=True | 689.933 (+-3.175) | 746.911 (+-2.985) | 2050.880 (+-22.188) | 2.746 (+-0.000) 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=False | | 78.347 (+-0.338) | 169.646 (+-1.640) | 2.165 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=False | | 162.194 (+-1.089) | 329.754 (+-2.590) | 2.033 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=False | | 894.476 (+-2.738) | 2815.870 (+-22.589) | 3.148 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=False | | 52.728 (+-0.406) | 112.024 (+-1.225) | 2.125 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=False | | 151.560 (+-1.128) | 299.152 (+-3.353) | 1.974 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=False | | 500.053 (+-4.288) | 1698.601 (+-16.785) | 3.397 (+-0.000) ``` Note: There is no perf regression for other case. There some cases (see Source below) with small speed-ups, for the rest it is roughly around 1.0 +/- 0.1 which may be attributed to noisy measurements ... [Source](https://gist.github.com/vfdev-5/1c0778904a07ce40401306548b9525e8#file-20230322-132441-pr_vs_nightly-speedup-md) ## Context - #90771 cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 [ghstack-poisoned]

…erpolate cpu uint8 RGB-case (channels last)" ## Description - Based on #96651 - Improved perfs for vectorized **bilinear** interpolate uint8 RGB-case, **channels last** - unified RGB and RGBA processing code such that RGB input is not copied into RGBA - Performances are more close to Pillow-SIMD (labeled as `Pillow (9.0.0.post1)` in the results) - RGBA case perfs are the same after refactoring (see Source link below) - Fixed mem pointer alignment, added more comments (reviews from #96651) ## Results - `Pillow (9.0.0.post1)` == Pillow-SIMD ``` [-------------------------------------------------------------------------------------------------- Resize -------------------------------------------------------------------------------------------------] | Pillow (9.0.0.post1) | torch (2.1.0a0+gitce4be01) PR | torch (2.1.0a0+git5309c44) nightly | Speed-up: PR vs nightly 1 threads: -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=True | 38.548 (+-0.280) | 57.536 (+-0.210) | 132.147 (+-1.236) | 2.297 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=False | | 38.532 (+-0.219) | 111.789 (+-1.175) | 2.901 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=True | 127.689 (+-1.348) | 156.262 (+-1.213) | 302.518 (+-2.632) | 1.936 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=False | | 145.483 (+-1.077) | 286.663 (+-2.494) | 1.970 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=True | 178.117 (+-1.956) | 215.053 (+-1.470) | 439.375 (+-4.014) | 2.043 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=False | | 211.340 (+-2.239) | 438.537 (+-4.143) | 2.075 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=True | 112.593 (+-1.266) | 130.414 (+-1.633) | 446.804 (+-3.283) | 3.426 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=False | | 58.767 (+-0.203) | 374.244 (+-13.598) | 6.368 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=True | 283.210 (+-2.937) | 324.157 (+-1.895) | 720.197 (+-3.467) | 2.222 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=False | | 239.800 (+-2.492) | 592.834 (+-3.903) | 2.472 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=True | 186.255 (+-1.629) | 204.834 (+-1.496) | 787.868 (+-3.648) | 3.846 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=False | | 77.335 (+-0.341) | 651.016 (+-3.926) | 8.418 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=True | 410.286 (+-2.439) | 443.934 (+-2.899) | 1123.923 (+-14.988) | 2.532 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=False | | 312.220 (+-2.307) | 915.347 (+-4.486) | 2.932 (+-0.000) # More test-cases from #90771 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=True | 60.611 (+-0.337) | 80.849 (+-1.780) | 170.465 (+-1.830) | 2.108 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=True | 132.971 (+-1.624) | 164.892 (+-1.426) | 330.971 (+-3.249) | 2.007 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=True | 948.467 (+-3.179) | 891.414 (+-5.282) | 2805.510 (+-25.503) | 3.147 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=True | 52.539 (+-0.327) | 72.471 (+-0.367) | 135.933 (+-1.625) | 1.876 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=True | 138.669 (+-1.867) | 168.628 (+-1.213) | 321.112 (+-2.904) | 1.904 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=True | 689.933 (+-3.175) | 746.911 (+-2.985) | 2050.880 (+-22.188) | 2.746 (+-0.000) 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=False | | 78.347 (+-0.338) | 169.646 (+-1.640) | 2.165 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=False | | 162.194 (+-1.089) | 329.754 (+-2.590) | 2.033 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=False | | 894.476 (+-2.738) | 2815.870 (+-22.589) | 3.148 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=False | | 52.728 (+-0.406) | 112.024 (+-1.225) | 2.125 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=False | | 151.560 (+-1.128) | 299.152 (+-3.353) | 1.974 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=False | | 500.053 (+-4.288) | 1698.601 (+-16.785) | 3.397 (+-0.000) ``` Note: There is no perf regression for other case. There some cases (see Source below) with small speed-ups, for the rest it is roughly around 1.0 +/- 0.1 which may be attributed to noisy measurements ... [Source](https://gist.github.com/vfdev-5/1c0778904a07ce40401306548b9525e8#file-20230322-132441-pr_vs_nightly-speedup-md) ## Context - #90771 cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 [ghstack-poisoned]

…t8 RGB-case (channels last)" ## Description - Based on #96651 - Improved perfs for vectorized **bilinear** interpolate uint8 RGB-case, **channels last** - unified RGB and RGBA processing code such that RGB input is not copied into RGBA - Performances are more close to Pillow-SIMD (labeled as `Pillow (9.0.0.post1)` in the results) - RGBA case perfs are the same after refactoring (see Source link below) - Fixed mem pointer alignment, added more comments (reviews from #96651) ## Results - `Pillow (9.0.0.post1)` == Pillow-SIMD ``` [-------------------------------------------------------------------------------------------------- Resize -------------------------------------------------------------------------------------------------] | Pillow (9.0.0.post1) | torch (2.1.0a0+gitce4be01) PR | torch (2.1.0a0+git5309c44) nightly | Speed-up: PR vs nightly 1 threads: -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=True | 38.548 (+-0.280) | 57.536 (+-0.210) | 132.147 (+-1.236) | 2.297 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=False | | 38.532 (+-0.219) | 111.789 (+-1.175) | 2.901 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=True | 127.689 (+-1.348) | 156.262 (+-1.213) | 302.518 (+-2.632) | 1.936 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=False | | 145.483 (+-1.077) | 286.663 (+-2.494) | 1.970 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=True | 178.117 (+-1.956) | 215.053 (+-1.470) | 439.375 (+-4.014) | 2.043 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=False | | 211.340 (+-2.239) | 438.537 (+-4.143) | 2.075 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=True | 112.593 (+-1.266) | 130.414 (+-1.633) | 446.804 (+-3.283) | 3.426 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=False | | 58.767 (+-0.203) | 374.244 (+-13.598) | 6.368 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=True | 283.210 (+-2.937) | 324.157 (+-1.895) | 720.197 (+-3.467) | 2.222 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=False | | 239.800 (+-2.492) | 592.834 (+-3.903) | 2.472 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=True | 186.255 (+-1.629) | 204.834 (+-1.496) | 787.868 (+-3.648) | 3.846 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=False | | 77.335 (+-0.341) | 651.016 (+-3.926) | 8.418 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=True | 410.286 (+-2.439) | 443.934 (+-2.899) | 1123.923 (+-14.988) | 2.532 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=False | | 312.220 (+-2.307) | 915.347 (+-4.486) | 2.932 (+-0.000) # More test-cases from #90771 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=True | 60.611 (+-0.337) | 80.849 (+-1.780) | 170.465 (+-1.830) | 2.108 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=True | 132.971 (+-1.624) | 164.892 (+-1.426) | 330.971 (+-3.249) | 2.007 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=True | 948.467 (+-3.179) | 891.414 (+-5.282) | 2805.510 (+-25.503) | 3.147 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=True | 52.539 (+-0.327) | 72.471 (+-0.367) | 135.933 (+-1.625) | 1.876 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=True | 138.669 (+-1.867) | 168.628 (+-1.213) | 321.112 (+-2.904) | 1.904 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=True | 689.933 (+-3.175) | 746.911 (+-2.985) | 2050.880 (+-22.188) | 2.746 (+-0.000) 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=False | | 78.347 (+-0.338) | 169.646 (+-1.640) | 2.165 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=False | | 162.194 (+-1.089) | 329.754 (+-2.590) | 2.033 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=False | | 894.476 (+-2.738) | 2815.870 (+-22.589) | 3.148 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=False | | 52.728 (+-0.406) | 112.024 (+-1.225) | 2.125 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=False | | 151.560 (+-1.128) | 299.152 (+-3.353) | 1.974 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=False | | 500.053 (+-4.288) | 1698.601 (+-16.785) | 3.397 (+-0.000) ``` Note: There is no perf regression for other case. There some cases (see Source below) with small speed-ups, for the rest it is roughly around 1.0 +/- 0.1 which may be attributed to noisy measurements ... [Source](https://gist.github.com/vfdev-5/1c0778904a07ce40401306548b9525e8#file-20230322-132441-pr_vs_nightly-speedup-md) ## Context - #90771 cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 [ghstack-poisoned]

…erpolate cpu uint8 RGB-case (channels last)" ## Description - Based on #96651 - Improved perfs for vectorized **bilinear** interpolate uint8 RGB-case, **channels last** - unified RGB and RGBA processing code such that RGB input is not copied into RGBA - Performances are more close to Pillow-SIMD (labeled as `Pillow (9.0.0.post1)` in the results) - RGBA case perfs are the same after refactoring (see Source link below) - Fixed mem pointer alignment, added more comments (reviews from #96651) ## Results - `Pillow (9.0.0.post1)` == Pillow-SIMD ``` [-------------------------------------------------------------------------------------------------- Resize -------------------------------------------------------------------------------------------------] | Pillow (9.0.0.post1) | torch (2.1.0a0+gitce4be01) PR | torch (2.1.0a0+git5309c44) nightly | Speed-up: PR vs nightly 1 threads: -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=True | 38.548 (+-0.280) | 57.536 (+-0.210) | 132.147 (+-1.236) | 2.297 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=False | | 38.532 (+-0.219) | 111.789 (+-1.175) | 2.901 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=True | 127.689 (+-1.348) | 156.262 (+-1.213) | 302.518 (+-2.632) | 1.936 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=False | | 145.483 (+-1.077) | 286.663 (+-2.494) | 1.970 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=True | 178.117 (+-1.956) | 215.053 (+-1.470) | 439.375 (+-4.014) | 2.043 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=False | | 211.340 (+-2.239) | 438.537 (+-4.143) | 2.075 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=True | 112.593 (+-1.266) | 130.414 (+-1.633) | 446.804 (+-3.283) | 3.426 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=False | | 58.767 (+-0.203) | 374.244 (+-13.598) | 6.368 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=True | 283.210 (+-2.937) | 324.157 (+-1.895) | 720.197 (+-3.467) | 2.222 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=False | | 239.800 (+-2.492) | 592.834 (+-3.903) | 2.472 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=True | 186.255 (+-1.629) | 204.834 (+-1.496) | 787.868 (+-3.648) | 3.846 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=False | | 77.335 (+-0.341) | 651.016 (+-3.926) | 8.418 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=True | 410.286 (+-2.439) | 443.934 (+-2.899) | 1123.923 (+-14.988) | 2.532 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=False | | 312.220 (+-2.307) | 915.347 (+-4.486) | 2.932 (+-0.000) # More test-cases from #90771 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=True | 60.611 (+-0.337) | 80.849 (+-1.780) | 170.465 (+-1.830) | 2.108 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=True | 132.971 (+-1.624) | 164.892 (+-1.426) | 330.971 (+-3.249) | 2.007 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=True | 948.467 (+-3.179) | 891.414 (+-5.282) | 2805.510 (+-25.503) | 3.147 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=True | 52.539 (+-0.327) | 72.471 (+-0.367) | 135.933 (+-1.625) | 1.876 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=True | 138.669 (+-1.867) | 168.628 (+-1.213) | 321.112 (+-2.904) | 1.904 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=True | 689.933 (+-3.175) | 746.911 (+-2.985) | 2050.880 (+-22.188) | 2.746 (+-0.000) 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=False | | 78.347 (+-0.338) | 169.646 (+-1.640) | 2.165 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=False | | 162.194 (+-1.089) | 329.754 (+-2.590) | 2.033 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=False | | 894.476 (+-2.738) | 2815.870 (+-22.589) | 3.148 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=False | | 52.728 (+-0.406) | 112.024 (+-1.225) | 2.125 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=False | | 151.560 (+-1.128) | 299.152 (+-3.353) | 1.974 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=False | | 500.053 (+-4.288) | 1698.601 (+-16.785) | 3.397 (+-0.000) ``` Note: There is no perf regression for other case. There some cases (see Source below) with small speed-ups, for the rest it is roughly around 1.0 +/- 0.1 which may be attributed to noisy measurements ... [Source](https://gist.github.com/vfdev-5/1c0778904a07ce40401306548b9525e8#file-20230322-132441-pr_vs_nightly-speedup-md) ## Context - #90771 cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 [ghstack-poisoned]

…t8 RGB-case (channels last)" ## Description - Based on #96651 - Improved perfs for vectorized **bilinear** interpolate uint8 RGB-case, **channels last** - unified RGB and RGBA processing code such that RGB input is not copied into RGBA - Performances are more close to Pillow-SIMD (labeled as `Pillow (9.0.0.post1)` in the results) - RGBA case perfs are the same after refactoring (see Source link below) - Fixed mem pointer alignment, added more comments (reviews from #96651) ## Results - `Pillow (9.0.0.post1)` == Pillow-SIMD ``` [-------------------------------------------------------------------------------------------------- Resize -------------------------------------------------------------------------------------------------] | Pillow (9.0.0.post1) | torch (2.1.0a0+gitce4be01) PR | torch (2.1.0a0+git5309c44) nightly | Speed-up: PR vs nightly 1 threads: -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=True | 38.548 (+-0.280) | 57.536 (+-0.210) | 132.147 (+-1.236) | 2.297 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=False | | 38.532 (+-0.219) | 111.789 (+-1.175) | 2.901 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=True | 127.689 (+-1.348) | 156.262 (+-1.213) | 302.518 (+-2.632) | 1.936 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=False | | 145.483 (+-1.077) | 286.663 (+-2.494) | 1.970 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=True | 178.117 (+-1.956) | 215.053 (+-1.470) | 439.375 (+-4.014) | 2.043 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=False | | 211.340 (+-2.239) | 438.537 (+-4.143) | 2.075 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=True | 112.593 (+-1.266) | 130.414 (+-1.633) | 446.804 (+-3.283) | 3.426 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=False | | 58.767 (+-0.203) | 374.244 (+-13.598) | 6.368 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=True | 283.210 (+-2.937) | 324.157 (+-1.895) | 720.197 (+-3.467) | 2.222 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=False | | 239.800 (+-2.492) | 592.834 (+-3.903) | 2.472 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=True | 186.255 (+-1.629) | 204.834 (+-1.496) | 787.868 (+-3.648) | 3.846 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=False | | 77.335 (+-0.341) | 651.016 (+-3.926) | 8.418 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=True | 410.286 (+-2.439) | 443.934 (+-2.899) | 1123.923 (+-14.988) | 2.532 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=False | | 312.220 (+-2.307) | 915.347 (+-4.486) | 2.932 (+-0.000) # More test-cases from #90771 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=True | 60.611 (+-0.337) | 80.849 (+-1.780) | 170.465 (+-1.830) | 2.108 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=True | 132.971 (+-1.624) | 164.892 (+-1.426) | 330.971 (+-3.249) | 2.007 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=True | 948.467 (+-3.179) | 891.414 (+-5.282) | 2805.510 (+-25.503) | 3.147 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=True | 52.539 (+-0.327) | 72.471 (+-0.367) | 135.933 (+-1.625) | 1.876 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=True | 138.669 (+-1.867) | 168.628 (+-1.213) | 321.112 (+-2.904) | 1.904 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=True | 689.933 (+-3.175) | 746.911 (+-2.985) | 2050.880 (+-22.188) | 2.746 (+-0.000) 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=False | | 78.347 (+-0.338) | 169.646 (+-1.640) | 2.165 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=False | | 162.194 (+-1.089) | 329.754 (+-2.590) | 2.033 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=False | | 894.476 (+-2.738) | 2815.870 (+-22.589) | 3.148 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=False | | 52.728 (+-0.406) | 112.024 (+-1.225) | 2.125 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=False | | 151.560 (+-1.128) | 299.152 (+-3.353) | 1.974 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=False | | 500.053 (+-4.288) | 1698.601 (+-16.785) | 3.397 (+-0.000) ``` Note: There is no perf regression for other case. There some cases (see Source below) with small speed-ups, for the rest it is roughly around 1.0 +/- 0.1 which may be attributed to noisy measurements ... [Source](https://gist.github.com/vfdev-5/1c0778904a07ce40401306548b9525e8#file-20230322-132441-pr_vs_nightly-speedup-md) ## Context - #90771 cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 [ghstack-poisoned]

…erpolate cpu uint8 RGB-case (channels last)" ## Description - Based on #96651 - Improved perfs for vectorized **bilinear** interpolate uint8 RGB-case, **channels last** - unified RGB and RGBA processing code such that RGB input is not copied into RGBA - Performances are more close to Pillow-SIMD (labeled as `Pillow (9.0.0.post1)` in the results) - RGBA case perfs are the same after refactoring (see Source link below) - Fixed mem pointer alignment, added more comments (reviews from #96651) ## Results - `Pillow (9.0.0.post1)` == Pillow-SIMD ``` [-------------------------------------------------------------------------------------------------- Resize -------------------------------------------------------------------------------------------------] | Pillow (9.0.0.post1) | torch (2.1.0a0+gitd6e220c) PR | torch (2.1.0a0+git2b75955) nightly | Speed-up: PR vs nightly 1 threads: -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=True | 38.674 (+-0.323) | 57.591 (+-0.244) | 131.033 (+-1.448) | 2.275 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=False | | 39.471 (+-0.166) | 113.911 (+-1.736) | 2.886 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=True | 128.512 (+-1.916) | 161.592 (+-1.242) | 299.679 (+-2.099) | 1.855 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=False | | 150.994 (+-1.180) | 285.331 (+-1.919) | 1.890 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=True | 180.045 (+-2.223) | 220.581 (+-1.363) | 431.057 (+-3.536) | 1.954 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=False | | 219.391 (+-1.409) | 429.410 (+-3.620) | 1.957 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=True | 113.911 (+-1.024) | 129.457 (+-1.295) | 459.610 (+-13.322) | 3.550 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=False | | 59.800 (+-0.199) | 400.015 (+-11.815) | 6.689 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=True | 283.050 (+-2.664) | 339.143 (+-1.209) | 683.555 (+-4.466) | 2.016 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=False | | 250.601 (+-1.236) | 603.545 (+-2.644) | 2.408 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=True | 186.723 (+-2.213) | 199.960 (+-1.343) | 860.867 (+-21.763) | 4.305 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=False | | 79.188 (+-0.261) | 703.019 (+-25.805) | 8.878 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=True | 412.353 (+-4.476) | 462.230 (+-1.983) | 1101.673 (+-49.299) | 2.383 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=False | | 327.973 (+-1.852) | 941.062 (+-5.549) | 2.869 (+-0.000) 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=True | 61.191 (+-0.926) | 80.795 (+-0.518) | 160.853 (+-1.506) | 1.991 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=True | 134.488 (+-2.129) | 169.147 (+-1.324) | 327.343 (+-2.846) | 1.935 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=True | 1037.045 (+-24.982) | 938.623 (+-9.010) | 2603.360 (+-20.530) | 2.774 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=True | 52.792 (+-0.613) | 73.692 (+-0.264) | 131.829 (+-1.333) | 1.789 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=True | 139.596 (+-1.944) | 173.778 (+-1.039) | 320.063 (+-2.562) | 1.842 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=True | 690.132 (+-10.946) | 772.758 (+-2.864) | 2036.860 (+-36.109) | 2.636 (+-0.000) 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=False | | 78.747 (+-0.799) | 158.479 (+-1.702) | 2.013 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=False | | 167.046 (+-1.077) | 322.104 (+-2.764) | 1.928 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=False | | 918.967 (+-5.251) | 2611.388 (+-29.917) | 2.842 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=False | | 55.336 (+-0.251) | 113.869 (+-1.243) | 2.058 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=False | | 156.505 (+-1.095) | 299.861 (+-2.710) | 1.916 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=False | | 514.344 (+-1.905) | 1776.796 (+-19.660) | 3.454 (+-0.000) ``` Note: There is no perf regression for other case. There some cases (see Source below) with small speed-ups, for the rest it is roughly around 1.0 +/- 0.1 which may be attributed to noisy measurements ... [Source](https://gist.github.com/vfdev-5/1c0778904a07ce40401306548b9525e8#file-20230329-181023-pr_vs_nightly-speedup-md) ## Context - #90771 cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 datumbox pmeier [ghstack-poisoned]

…t8 RGB-case (channels last)" ## Description - Based on #96651 - Improved perfs for vectorized **bilinear** interpolate uint8 RGB-case, **channels last** - unified RGB and RGBA processing code such that RGB input is not copied into RGBA - Performances are more close to Pillow-SIMD (labeled as `Pillow (9.0.0.post1)` in the results) - RGBA case perfs are the same after refactoring (see Source link below) - Fixed mem pointer alignment, added more comments (reviews from #96651) ## Results - `Pillow (9.0.0.post1)` == Pillow-SIMD ``` [-------------------------------------------------------------------------------------------------- Resize -------------------------------------------------------------------------------------------------] | Pillow (9.0.0.post1) | torch (2.1.0a0+gitd6e220c) PR | torch (2.1.0a0+git2b75955) nightly | Speed-up: PR vs nightly 1 threads: -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=True | 38.674 (+-0.323) | 57.591 (+-0.244) | 131.033 (+-1.448) | 2.275 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=False | | 39.471 (+-0.166) | 113.911 (+-1.736) | 2.886 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=True | 128.512 (+-1.916) | 161.592 (+-1.242) | 299.679 (+-2.099) | 1.855 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=False | | 150.994 (+-1.180) | 285.331 (+-1.919) | 1.890 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=True | 180.045 (+-2.223) | 220.581 (+-1.363) | 431.057 (+-3.536) | 1.954 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=False | | 219.391 (+-1.409) | 429.410 (+-3.620) | 1.957 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=True | 113.911 (+-1.024) | 129.457 (+-1.295) | 459.610 (+-13.322) | 3.550 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=False | | 59.800 (+-0.199) | 400.015 (+-11.815) | 6.689 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=True | 283.050 (+-2.664) | 339.143 (+-1.209) | 683.555 (+-4.466) | 2.016 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=False | | 250.601 (+-1.236) | 603.545 (+-2.644) | 2.408 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=True | 186.723 (+-2.213) | 199.960 (+-1.343) | 860.867 (+-21.763) | 4.305 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=False | | 79.188 (+-0.261) | 703.019 (+-25.805) | 8.878 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=True | 412.353 (+-4.476) | 462.230 (+-1.983) | 1101.673 (+-49.299) | 2.383 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=False | | 327.973 (+-1.852) | 941.062 (+-5.549) | 2.869 (+-0.000) 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=True | 61.191 (+-0.926) | 80.795 (+-0.518) | 160.853 (+-1.506) | 1.991 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=True | 134.488 (+-2.129) | 169.147 (+-1.324) | 327.343 (+-2.846) | 1.935 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=True | 1037.045 (+-24.982) | 938.623 (+-9.010) | 2603.360 (+-20.530) | 2.774 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=True | 52.792 (+-0.613) | 73.692 (+-0.264) | 131.829 (+-1.333) | 1.789 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=True | 139.596 (+-1.944) | 173.778 (+-1.039) | 320.063 (+-2.562) | 1.842 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=True | 690.132 (+-10.946) | 772.758 (+-2.864) | 2036.860 (+-36.109) | 2.636 (+-0.000) 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=False | | 78.747 (+-0.799) | 158.479 (+-1.702) | 2.013 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=False | | 167.046 (+-1.077) | 322.104 (+-2.764) | 1.928 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=False | | 918.967 (+-5.251) | 2611.388 (+-29.917) | 2.842 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=False | | 55.336 (+-0.251) | 113.869 (+-1.243) | 2.058 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=False | | 156.505 (+-1.095) | 299.861 (+-2.710) | 1.916 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=False | | 514.344 (+-1.905) | 1776.796 (+-19.660) | 3.454 (+-0.000) ``` Note: There is no perf regression for other case. There some cases (see Source below) with small speed-ups, for the rest it is roughly around 1.0 +/- 0.1 which may be attributed to noisy measurements ... [Source](https://gist.github.com/vfdev-5/1c0778904a07ce40401306548b9525e8#file-20230329-181023-pr_vs_nightly-speedup-md) ## Context - #90771 cc jgong5 mingfeima XiaobingSuper sanchitintel ashokei jingxu10 datumbox pmeier [ghstack-poisoned]

… (channels last) (#96848) ## Description - Based on #96651 - Improved perfs for vectorized **bilinear** interpolate uint8 RGB-case, **channels last** - unified RGB and RGBA processing code such that RGB input is not copied into RGBA - Performances are more close to Pillow-SIMD (labeled as `Pillow (9.0.0.post1)` in the results) - RGBA case perfs are the same after refactoring (see Source link below) - Fixed mem pointer alignment, added more comments (reviews from #96651) ## Results - `Pillow (9.0.0.post1)` == Pillow-SIMD ``` [-------------------------------------------------------------------------------------------------- Resize -------------------------------------------------------------------------------------------------] | Pillow (9.0.0.post1) | torch (2.1.0a0+gitd6e220c) PR | torch (2.1.0a0+git2b75955) nightly | Speed-up: PR vs nightly 1 threads: -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=True | 38.674 (+-0.323) | 57.591 (+-0.244) | 131.033 (+-1.448) | 2.275 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=False | | 39.471 (+-0.166) | 113.911 (+-1.736) | 2.886 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=True | 128.512 (+-1.916) | 161.592 (+-1.242) | 299.679 (+-2.099) | 1.855 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=False | | 150.994 (+-1.180) | 285.331 (+-1.919) | 1.890 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=True | 180.045 (+-2.223) | 220.581 (+-1.363) | 431.057 (+-3.536) | 1.954 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=False | | 219.391 (+-1.409) | 429.410 (+-3.620) | 1.957 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=True | 113.911 (+-1.024) | 129.457 (+-1.295) | 459.610 (+-13.322) | 3.550 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=False | | 59.800 (+-0.199) | 400.015 (+-11.815) | 6.689 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=True | 283.050 (+-2.664) | 339.143 (+-1.209) | 683.555 (+-4.466) | 2.016 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=False | | 250.601 (+-1.236) | 603.545 (+-2.644) | 2.408 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=True | 186.723 (+-2.213) | 199.960 (+-1.343) | 860.867 (+-21.763) | 4.305 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=False | | 79.188 (+-0.261) | 703.019 (+-25.805) | 8.878 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=True | 412.353 (+-4.476) | 462.230 (+-1.983) | 1101.673 (+-49.299) | 2.383 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=False | | 327.973 (+-1.852) | 941.062 (+-5.549) | 2.869 (+-0.000) 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=True | 61.191 (+-0.926) | 80.795 (+-0.518) | 160.853 (+-1.506) | 1.991 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=True | 134.488 (+-2.129) | 169.147 (+-1.324) | 327.343 (+-2.846) | 1.935 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=True | 1037.045 (+-24.982) | 938.623 (+-9.010) | 2603.360 (+-20.530) | 2.774 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=True | 52.792 (+-0.613) | 73.692 (+-0.264) | 131.829 (+-1.333) | 1.789 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=True | 139.596 (+-1.944) | 173.778 (+-1.039) | 320.063 (+-2.562) | 1.842 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=True | 690.132 (+-10.946) | 772.758 (+-2.864) | 2036.860 (+-36.109) | 2.636 (+-0.000) 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=False | | 78.747 (+-0.799) | 158.479 (+-1.702) | 2.013 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=False | | 167.046 (+-1.077) | 322.104 (+-2.764) | 1.928 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=False | | 918.967 (+-5.251) | 2611.388 (+-29.917) | 2.842 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=False | | 55.336 (+-0.251) | 113.869 (+-1.243) | 2.058 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=False | | 156.505 (+-1.095) | 299.861 (+-2.710) | 1.916 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=False | | 514.344 (+-1.905) | 1776.796 (+-19.660) | 3.454 (+-0.000) ``` Note: There is no perf regression for other case. There some cases (see Source below) with small speed-ups, for the rest it is roughly around 1.0 +/- 0.1 which may be attributed to noisy measurements ... [Source](https://gist.github.com/vfdev-5/1c0778904a07ce40401306548b9525e8#file-20230329-181023-pr_vs_nightly-speedup-md) ## Context - #90771 Pull Request resolved: #96848 Approved by: https://github.com/NicolasHug, https://github.com/peterbell10

Description: - Now that pytorch/pytorch#90771 is merged, let Resize() rely on interpolate()'s native uint8 handling instead of converting to and from float. - uint8 input is not casted to f32 for nearest mode and bilinear mode if the latter has AVX2. Context: pytorch#7217 Benchmarks: ``` [----------- Resize cpu torch.uint8 InterpolationMode.NEAREST -----------] | resize v2 | resize stable | resize nightly 1 threads: --------------------------------------------------------------- (3, 400, 400) | 457 | 461 | 480 (16, 3, 400, 400) | 6870 | 6850 | 10100 Times are in microseconds (us). [---------- Resize cpu torch.uint8 InterpolationMode.BILINEAR -----------] | resize v2 | resize stable | resize nightly 1 threads: --------------------------------------------------------------- (3, 400, 400) | 326 | 329 | 844 (16, 3, 400, 400) | 4380 | 4390 | 14800 Times are in microseconds (us). ``` [Source](https://gist.github.com/vfdev-5/a2e30ed50b5996807c9b09d5d33d8bc2)

@vfdev-5

Joint work with @vfdev-5 This PR introduces native uint8 support for `interpolate()`, for `bilinear` ~and `bicubic`~ modes for CPU images (`mode=nearest[_exact]` was already supported ). On a typical torchvision training job on ImageNet, the speedup are ~4X when AVX2 is supported, comparing the uint8 native (this PR) vs torchvision's current `Resize()`: ``` AA = antialias float = uint8->float->interpolate()->round()->clamp()->uint8 (what Resize() currently does) input_size output_size channels_last AA mode num_threads speed-up float vs uint8 (this PR) (1, 3, 270, 268) -> (224, 224) True True bilinear num_threads=1 4X 2.6ms vs 0.7ms (1, 3, 270, 268) -> (224, 224) True False bilinear num_threads=1 2.1X 1.3ms vs 0.6ms (1, 3, 270, 268) -> (224, 224) False True bilinear num_threads=1 3X 2.1ms vs 0.7ms (1, 3, 270, 268) -> (224, 224) False False bilinear num_threads=1 4X 2.4ms vs 0.6ms (Note: we removed bicubic support for now) (1, 3, 270, 268) -> (224, 224) True True bicubic num_threads=1 4X 2.9ms vs 0.7ms (1, 3, 270, 268) -> (224, 224) True False bicubic num_threads=1 5X 3.1ms vs 0.7ms (1, 3, 270, 268) -> (224, 224) False True bicubic num_threads=1 3X 2.4ms vs 0.7ms (1, 3, 270, 268) -> (224, 224) False False bicubic num_threads=1 4X 2.8ms vs 0.7ms ``` There is still room for further speed-ups (see TODOs in the code). #### More benchmark details with AVX2 support - speedups typically range from 1.5X to 10X. A few edge-cases are slower, worth investigating why. <details> ``` AA = antialias float = uint8->float->interpolate()->round()->clamp()->uint8 (what Resize() currently does) input_size output_size channels_last AA mode num_threads speed-up float vs uint8 (this PR) (1, 3, 64, 64) -> (224, 224) True True bilinear num_threads=1 5X 1.1ms vs 0.2ms (1, 3, 64, 64) -> (224, 224) True False bilinear num_threads=1 5X 1.2ms vs 0.2ms (1, 3, 64, 64) -> (224, 224) False True bilinear num_threads=1 2.8X 0.6ms vs 0.2ms (1, 3, 64, 64) -> (224, 224) False False bilinear num_threads=1 7X 1.6ms vs 0.2ms (1, 3, 64, 64) -> (224, 224) True True bicubic num_threads=1 5X 1.2ms vs 0.2ms (1, 3, 64, 64) -> (224, 224) True False bicubic num_threads=1 12X 2.9ms vs 0.2ms (1, 3, 64, 64) -> (224, 224) False True bicubic num_threads=1 3X 0.8ms vs 0.2ms (1, 3, 64, 64) -> (224, 224) False False bicubic num_threads=1 7X 1.8ms vs 0.2ms (1, 3, 64, 64) -> (224, 224) True True bilinear num_threads=2 2.6X 0.6ms vs 0.2ms (1, 3, 64, 64) -> (224, 224) True False bilinear num_threads=2 2.8X 0.6ms vs 0.2ms (1, 3, 64, 64) -> (224, 224) False True bilinear num_threads=2 1.7X 0.4ms vs 0.2ms (1, 3, 64, 64) -> (224, 224) False False bilinear num_threads=2 1.4X 0.3ms vs 0.2ms (1, 3, 64, 64) -> (224, 224) True True bicubic num_threads=2 2.7X 0.7ms vs 0.2ms (1, 3, 64, 64) -> (224, 224) True False bicubic num_threads=2 7X 1.6ms vs 0.2ms (1, 3, 64, 64) -> (224, 224) False True bicubic num_threads=2 1.8X 0.4ms vs 0.2ms (1, 3, 64, 64) -> (224, 224) False False bicubic num_threads=2 4X 1.0ms vs 0.2ms (1, 3, 224, 224) -> (270, 268) True True bilinear num_threads=1 4X 2.5ms vs 0.6ms (1, 3, 224, 224) -> (270, 268) True False bilinear num_threads=1 3.0X 1.8ms vs 0.6ms (1, 3, 224, 224) -> (270, 268) False True bilinear num_threads=1 3X 1.8ms vs 0.6ms (1, 3, 224, 224) -> (270, 268) False False bilinear num_threads=1 4X 2.3ms vs 0.6ms (1, 3, 224, 224) -> (270, 268) True True bicubic num_threads=1 4X 2.7ms vs 0.6ms (1, 3, 224, 224) -> (270, 268) True False bicubic num_threads=1 7X 4.3ms vs 0.6ms (1, 3, 224, 224) -> (270, 268) False True bicubic num_threads=1 3X 2.1ms vs 0.6ms (1, 3, 224, 224) -> (270, 268) False False bicubic num_threads=1 4X 2.6ms vs 0.6ms (1, 3, 224, 224) -> (270, 268) True True bilinear num_threads=2 2.7X 1.6ms vs 0.6ms (1, 3, 224, 224) -> (270, 268) True False bilinear num_threads=2 2.6X 1.5ms vs 0.6ms (1, 3, 224, 224) -> (270, 268) False True bilinear num_threads=2 2.1X 1.2ms vs 0.6ms (1, 3, 224, 224) -> (270, 268) False False bilinear num_threads=2 1.6X 0.9ms vs 0.6ms (1, 3, 224, 224) -> (270, 268) True True bicubic num_threads=2 2.8X 1.7ms vs 0.6ms (1, 3, 224, 224) -> (270, 268) True False bicubic num_threads=2 5X 2.8ms vs 0.6ms (1, 3, 224, 224) -> (270, 268) False True bicubic num_threads=2 2.3X 1.4ms vs 0.6ms (1, 3, 224, 224) -> (270, 268) False False bicubic num_threads=2 3X 1.9ms vs 0.6ms (1, 3, 256, 256) -> (1024, 1024) True True bilinear num_threads=1 4X 26.6ms vs 6.7ms (1, 3, 256, 256) -> (1024, 1024) True False bilinear num_threads=1 4X 23.9ms vs 6.8ms (1, 3, 256, 256) -> (1024, 1024) False True bilinear num_threads=1 2.5X 16.8ms vs 6.8ms (1, 3, 256, 256) -> (1024, 1024) False False bilinear num_threads=1 5X 33.1ms vs 6.8ms (1, 3, 256, 256) -> (1024, 1024) True True bicubic num_threads=1 4X 25.9ms vs 7.3ms (1, 3, 256, 256) -> (1024, 1024) True False bicubic num_threads=1 8X 59.6ms vs 7.3ms (1, 3, 256, 256) -> (1024, 1024) False True bicubic num_threads=1 1.9X 14.3ms vs 7.4ms (1, 3, 256, 256) -> (1024, 1024) False False bicubic num_threads=1 5X 35.4ms vs 7.3ms (1, 3, 256, 256) -> (1024, 1024) True True bilinear num_threads=2 2.0X 13.6ms vs 6.8ms (1, 3, 256, 256) -> (1024, 1024) True False bilinear num_threads=2 2.2X 14.8ms vs 6.7ms (1, 3, 256, 256) -> (1024, 1024) False True bilinear num_threads=2 1.3X 8.8ms vs 6.9ms (1, 3, 256, 256) -> (1024, 1024) False False bilinear num_threads=2 1.2X 8.4ms vs 6.8ms (1, 3, 256, 256) -> (1024, 1024) True True bicubic num_threads=2 1.8X 12.8ms vs 7.3ms (1, 3, 256, 256) -> (1024, 1024) True False bicubic num_threads=2 4X 32.1ms vs 7.2ms (1, 3, 256, 256) -> (1024, 1024) False True bicubic num_threads=2 1.4X 10.1ms vs 7.3ms (1, 3, 256, 256) -> (1024, 1024) False False bicubic num_threads=2 2.9X 20.9ms vs 7.3ms (1, 3, 224, 224) -> (64, 64) True True bilinear num_threads=1 1.4X 0.5ms vs 0.3ms (1, 3, 224, 224) -> (64, 64) True False bilinear num_threads=1 0.7X 0.2ms vs 0.3ms (1, 3, 224, 224) -> (64, 64) False True bilinear num_threads=1 1.3X 0.4ms vs 0.3ms (1, 3, 224, 224) -> (64, 64) False False bilinear num_threads=1 1.4X 0.4ms vs 0.3ms (1, 3, 224, 224) -> (64, 64) True True bicubic num_threads=1 2.1X 0.7ms vs 0.3ms (1, 3, 224, 224) -> (64, 64) True False bicubic num_threads=1 1.3X 0.4ms vs 0.3ms (1, 3, 224, 224) -> (64, 64) False True bicubic num_threads=1 1.9X 0.6ms vs 0.3ms (1, 3, 224, 224) -> (64, 64) False False bicubic num_threads=1 1.0X 0.3ms vs 0.3ms (1, 3, 224, 224) -> (64, 64) True True bilinear num_threads=2 1.0X 0.3ms vs 0.3ms (1, 3, 224, 224) -> (64, 64) True False bilinear num_threads=2 0.6X 0.2ms vs 0.3ms (1, 3, 224, 224) -> (64, 64) False True bilinear num_threads=2 0.8X 0.3ms vs 0.3ms (1, 3, 224, 224) -> (64, 64) False False bilinear num_threads=2 1.4X 0.4ms vs 0.3ms (1, 3, 224, 224) -> (64, 64) True True bicubic num_threads=2 1.4X 0.5ms vs 0.3ms (1, 3, 224, 224) -> (64, 64) True False bicubic num_threads=2 1.2X 0.4ms vs 0.3ms (1, 3, 224, 224) -> (64, 64) False True bicubic num_threads=2 1.2X 0.4ms vs 0.4ms (1, 3, 224, 224) -> (64, 64) False False bicubic num_threads=2 0.9X 0.3ms vs 0.3ms (1, 3, 270, 268) -> (224, 224) True True bilinear num_threads=1 4X 2.6ms vs 0.7ms (1, 3, 270, 268) -> (224, 224) True False bilinear num_threads=1 2.1X 1.3ms vs 0.6ms (1, 3, 270, 268) -> (224, 224) False True bilinear num_threads=1 3X 2.1ms vs 0.7ms (1, 3, 270, 268) -> (224, 224) False False bilinear num_threads=1 4X 2.4ms vs 0.6ms (1, 3, 270, 268) -> (224, 224) True True bicubic num_threads=1 4X 2.9ms vs 0.7ms (1, 3, 270, 268) -> (224, 224) True False bicubic num_threads=1 5X 3.1ms vs 0.7ms (1, 3, 270, 268) -> (224, 224) False True bicubic num_threads=1 3X 2.4ms vs 0.7ms (1, 3, 270, 268) -> (224, 224) False False bicubic num_threads=1 4X 2.8ms vs 0.7ms (1, 3, 270, 268) -> (224, 224) True True bilinear num_threads=2 1.5X 1.0ms vs 0.7ms (1, 3, 270, 268) -> (224, 224) True False bilinear num_threads=2 1.2X 0.8ms vs 0.6ms (1, 3, 270, 268) -> (224, 224) False True bilinear num_threads=2 2.3X 1.5ms vs 0.7ms (1, 3, 270, 268) -> (224, 224) False False bilinear num_threads=2 1.9X 1.2ms vs 0.6ms (1, 3, 270, 268) -> (224, 224) True True bicubic num_threads=2 1.6X 1.2ms vs 0.7ms (1, 3, 270, 268) -> (224, 224) True False bicubic num_threads=2 4X 2.4ms vs 0.7ms (1, 3, 270, 268) -> (224, 224) False True bicubic num_threads=2 2.4X 1.6ms vs 0.7ms (1, 3, 270, 268) -> (224, 224) False False bicubic num_threads=2 2.8X 1.8ms vs 0.6ms (1, 3, 1024, 1024) -> (256, 256) True True bilinear num_threads=1 2.1X 12.8ms vs 6.1ms (1, 3, 1024, 1024) -> (256, 256) True False bilinear num_threads=1 0.6X 3.8ms vs 5.9ms (1, 3, 1024, 1024) -> (256, 256) False True bilinear num_threads=1 1.2X 7.1ms vs 6.1ms (1, 3, 1024, 1024) -> (256, 256) False False bilinear num_threads=1 1.9X 11.0ms vs 5.9ms (1, 3, 1024, 1024) -> (256, 256) True True bicubic num_threads=1 2.0X 12.6ms vs 6.4ms (1, 3, 1024, 1024) -> (256, 256) True False bicubic num_threads=1 1.0X 6.1ms vs 6.0ms (1, 3, 1024, 1024) -> (256, 256) False True bicubic num_threads=1 1.8X 11.3ms vs 6.4ms (1, 3, 1024, 1024) -> (256, 256) False False bicubic num_threads=1 0.8X 4.6ms vs 6.0ms (1, 3, 1024, 1024) -> (256, 256) True True bilinear num_threads=2 1.6X 9.3ms vs 6.0ms (1, 3, 1024, 1024) -> (256, 256) True False bilinear num_threads=2 0.3X 2.0ms vs 5.8ms (1, 3, 1024, 1024) -> (256, 256) False True bilinear num_threads=2 1.2X 7.2ms vs 6.0ms (1, 3, 1024, 1024) -> (256, 256) False False bilinear num_threads=2 0.3X 1.6ms vs 5.8ms (1, 3, 1024, 1024) -> (256, 256) True True bicubic num_threads=2 1.1X 7.1ms vs 6.5ms (1, 3, 1024, 1024) -> (256, 256) True False bicubic num_threads=2 0.6X 3.3ms vs 5.9ms (1, 3, 1024, 1024) -> (256, 256) False True bicubic num_threads=2 0.9X 5.9ms vs 6.3ms (1, 3, 1024, 1024) -> (256, 256) False False bicubic num_threads=2 0.4X 2.4ms vs 5.9ms ``` </details> without AVX2 support - no significant speed-up, but there are various possible improvements (see TODOs) <details> ``` AA = antialias float = uint8->float->interpolate()->round()->clamp()->uint8 (what Resize() currently does) input_size output_size channels_last AA mode num_threads speed-up float vs uint8 (this PR) (1, 3, 64, 64) -> (224, 224) True True bilinear num_threads=1 0.9X 1.5ms vs 1.6ms (1, 3, 64, 64) -> (224, 224) True False bilinear num_threads=1 0.9X 1.5ms vs 1.6ms (1, 3, 64, 64) -> (224, 224) False True bilinear num_threads=1 0.8X 0.9ms vs 1.1ms (1, 3, 64, 64) -> (224, 224) False False bilinear num_threads=1 1.5X 1.7ms vs 1.1ms (1, 3, 64, 64) -> (224, 224) True True bicubic num_threads=1 0.9X 1.6ms vs 1.8ms (1, 3, 64, 64) -> (224, 224) True False bicubic num_threads=1 2.1X 3.9ms vs 1.9ms (1, 3, 64, 64) -> (224, 224) False True bicubic num_threads=1 0.8X 1.1ms vs 1.4ms (1, 3, 64, 64) -> (224, 224) False False bicubic num_threads=1 1.7X 2.4ms vs 1.5ms (1, 3, 64, 64) -> (224, 224) True True bilinear num_threads=2 0.9X 0.8ms vs 0.8ms (1, 3, 64, 64) -> (224, 224) True False bilinear num_threads=2 0.9X 0.8ms vs 0.8ms (1, 3, 64, 64) -> (224, 224) False True bilinear num_threads=2 0.9X 0.5ms vs 0.6ms (1, 3, 64, 64) -> (224, 224) False False bilinear num_threads=2 0.7X 0.5ms vs 0.7ms (1, 3, 64, 64) -> (224, 224) True True bicubic num_threads=2 0.9X 0.9ms vs 1.0ms (1, 3, 64, 64) -> (224, 224) True False bicubic num_threads=2 2.1X 2.0ms vs 1.0ms (1, 3, 64, 64) -> (224, 224) False True bicubic num_threads=2 0.8X 0.6ms vs 0.8ms (1, 3, 64, 64) -> (224, 224) False False bicubic num_threads=2 1.7X 1.3ms vs 0.8ms (1, 3, 224, 224) -> (270, 268) True True bilinear num_threads=1 1.0X 3.0ms vs 3.0ms (1, 3, 224, 224) -> (270, 268) True False bilinear num_threads=1 1.0X 2.8ms vs 2.9ms (1, 3, 224, 224) -> (270, 268) False True bilinear num_threads=1 1.0X 2.3ms vs 2.2ms (1, 3, 224, 224) -> (270, 268) False False bilinear num_threads=1 1.4X 3.3ms vs 2.3ms (1, 3, 224, 224) -> (270, 268) True True bicubic num_threads=1 1.0X 3.5ms vs 3.5ms (1, 3, 224, 224) -> (270, 268) True False bicubic num_threads=1 1.7X 6.1ms vs 3.5ms (1, 3, 224, 224) -> (270, 268) False True bicubic num_threads=1 0.9X 2.6ms vs 2.9ms (1, 3, 224, 224) -> (270, 268) False False bicubic num_threads=1 1.4X 4.2ms vs 2.9ms (1, 3, 224, 224) -> (270, 268) True True bilinear num_threads=2 1.0X 1.7ms vs 1.7ms (1, 3, 224, 224) -> (270, 268) True False bilinear num_threads=2 0.9X 1.6ms vs 1.8ms (1, 3, 224, 224) -> (270, 268) False True bilinear num_threads=2 0.9X 1.3ms vs 1.4ms (1, 3, 224, 224) -> (270, 268) False False bilinear num_threads=2 0.7X 1.1ms vs 1.6ms (1, 3, 224, 224) -> (270, 268) True True bicubic num_threads=2 1.0X 2.0ms vs 2.0ms (1, 3, 224, 224) -> (270, 268) True False bicubic num_threads=2 1.7X 3.2ms vs 1.9ms (1, 3, 224, 224) -> (270, 268) False True bicubic num_threads=2 0.8X 1.5ms vs 1.9ms (1, 3, 224, 224) -> (270, 268) False False bicubic num_threads=2 1.2X 2.3ms vs 1.9ms (1, 3, 256, 256) -> (1024, 1024) True True bilinear num_threads=1 1.1X 34.7ms vs 32.4ms (1, 3, 256, 256) -> (1024, 1024) True False bilinear num_threads=1 1.0X 31.2ms vs 32.4ms (1, 3, 256, 256) -> (1024, 1024) False True bilinear num_threads=1 1.0X 23.5ms vs 22.7ms (1, 3, 256, 256) -> (1024, 1024) False False bilinear num_threads=1 1.9X 42.5ms vs 22.7ms (1, 3, 256, 256) -> (1024, 1024) True True bicubic num_threads=1 0.9X 33.9ms vs 37.4ms (1, 3, 256, 256) -> (1024, 1024) True False bicubic num_threads=1 2.2X 84.0ms vs 37.5ms (1, 3, 256, 256) -> (1024, 1024) False True bicubic num_threads=1 1.0X 28.4ms vs 28.8ms (1, 3, 256, 256) -> (1024, 1024) False False bicubic num_threads=1 2.0X 56.7ms vs 28.8ms (1, 3, 256, 256) -> (1024, 1024) True True bilinear num_threads=2 1.1X 17.5ms vs 16.4ms (1, 3, 256, 256) -> (1024, 1024) True False bilinear num_threads=2 1.1X 17.7ms vs 16.4ms (1, 3, 256, 256) -> (1024, 1024) False True bilinear num_threads=2 0.8X 8.8ms vs 11.4ms (1, 3, 256, 256) -> (1024, 1024) False False bilinear num_threads=2 1.0X 11.1ms vs 11.4ms (1, 3, 256, 256) -> (1024, 1024) True True bicubic num_threads=2 1.1X 19.9ms vs 18.8ms (1, 3, 256, 256) -> (1024, 1024) True False bicubic num_threads=2 2.3X 42.5ms vs 18.7ms (1, 3, 256, 256) -> (1024, 1024) False True bicubic num_threads=2 1.0X 14.1ms vs 14.5ms (1, 3, 256, 256) -> (1024, 1024) False False bicubic num_threads=2 2.0X 28.4ms vs 14.5ms (1, 3, 224, 224) -> (64, 64) True True bilinear num_threads=1 1.0X 0.6ms vs 0.6ms (1, 3, 224, 224) -> (64, 64) True False bilinear num_threads=1 0.7X 0.3ms vs 0.4ms (1, 3, 224, 224) -> (64, 64) False True bilinear num_threads=1 0.9X 0.5ms vs 0.6ms (1, 3, 224, 224) -> (64, 64) False False bilinear num_threads=1 1.7X 0.6ms vs 0.4ms (1, 3, 224, 224) -> (64, 64) True True bicubic num_threads=1 1.0X 0.8ms vs 0.8ms (1, 3, 224, 224) -> (64, 64) True False bicubic num_threads=1 1.1X 0.5ms vs 0.5ms (1, 3, 224, 224) -> (64, 64) False True bicubic num_threads=1 0.9X 0.7ms vs 0.8ms (1, 3, 224, 224) -> (64, 64) False False bicubic num_threads=1 0.9X 0.4ms vs 0.4ms (1, 3, 224, 224) -> (64, 64) True True bilinear num_threads=2 1.0X 0.4ms vs 0.4ms (1, 3, 224, 224) -> (64, 64) True False bilinear num_threads=2 0.8X 0.2ms vs 0.3ms (1, 3, 224, 224) -> (64, 64) False True bilinear num_threads=2 0.9X 0.3ms vs 0.3ms (1, 3, 224, 224) -> (64, 64) False False bilinear num_threads=2 1.3X 0.3ms vs 0.2ms (1, 3, 224, 224) -> (64, 64) True True bicubic num_threads=2 1.0X 0.5ms vs 0.5ms (1, 3, 224, 224) -> (64, 64) True False bicubic num_threads=2 1.3X 0.4ms vs 0.3ms (1, 3, 224, 224) -> (64, 64) False True bicubic num_threads=2 0.9X 0.5ms vs 0.5ms (1, 3, 224, 224) -> (64, 64) False False bicubic num_threads=2 1.2X 0.3ms vs 0.3ms (1, 3, 270, 268) -> (224, 224) True True bilinear num_threads=1 0.8X 2.1ms vs 2.5ms (1, 3, 270, 268) -> (224, 224) True False bilinear num_threads=1 0.7X 1.6ms vs 2.4ms (1, 3, 270, 268) -> (224, 224) False True bilinear num_threads=1 1.2X 2.4ms vs 2.1ms (1, 3, 270, 268) -> (224, 224) False False bilinear num_threads=1 1.3X 2.6ms vs 2.0ms (1, 3, 270, 268) -> (224, 224) True True bicubic num_threads=1 1.1X 3.4ms vs 3.0ms (1, 3, 270, 268) -> (224, 224) True False bicubic num_threads=1 1.7X 4.8ms vs 2.8ms (1, 3, 270, 268) -> (224, 224) False True bicubic num_threads=1 1.1X 2.9ms vs 2.7ms (1, 3, 270, 268) -> (224, 224) False False bicubic num_threads=1 1.4X 3.5ms vs 2.4ms (1, 3, 270, 268) -> (224, 224) True True bilinear num_threads=2 0.9X 1.2ms vs 1.3ms (1, 3, 270, 268) -> (224, 224) True False bilinear num_threads=2 1.3X 1.6ms vs 1.2ms (1, 3, 270, 268) -> (224, 224) False True bilinear num_threads=2 0.8X 0.9ms vs 1.1ms (1, 3, 270, 268) -> (224, 224) False False bilinear num_threads=2 1.3X 1.3ms vs 1.0ms (1, 3, 270, 268) -> (224, 224) True True bicubic num_threads=2 1.4X 2.2ms vs 1.6ms (1, 3, 270, 268) -> (224, 224) True False bicubic num_threads=2 1.9X 2.8ms vs 1.5ms (1, 3, 270, 268) -> (224, 224) False True bicubic num_threads=2 0.8X 1.1ms vs 1.4ms (1, 3, 270, 268) -> (224, 224) False False bicubic num_threads=2 1.7X 2.1ms vs 1.3ms (1, 3, 1024, 1024) -> (256, 256) True True bilinear num_threads=1 1.0X 10.0ms vs 9.9ms (1, 3, 1024, 1024) -> (256, 256) True False bilinear num_threads=1 0.7X 4.6ms vs 6.2ms (1, 3, 1024, 1024) -> (256, 256) False True bilinear num_threads=1 0.9X 9.1ms vs 9.8ms (1, 3, 1024, 1024) -> (256, 256) False False bilinear num_threads=1 1.7X 9.4ms vs 5.7ms (1, 3, 1024, 1024) -> (256, 256) True True bicubic num_threads=1 1.0X 15.2ms vs 14.8ms (1, 3, 1024, 1024) -> (256, 256) True False bicubic num_threads=1 1.0X 7.6ms vs 7.5ms (1, 3, 1024, 1024) -> (256, 256) False True bicubic num_threads=1 0.9X 13.3ms vs 14.4ms (1, 3, 1024, 1024) -> (256, 256) False False bicubic num_threads=1 0.8X 5.9ms vs 7.0ms (1, 3, 1024, 1024) -> (256, 256) True True bilinear num_threads=2 1.2X 6.0ms vs 5.2ms (1, 3, 1024, 1024) -> (256, 256) True False bilinear num_threads=2 0.7X 2.3ms vs 3.2ms (1, 3, 1024, 1024) -> (256, 256) False True bilinear num_threads=2 1.0X 4.8ms vs 5.0ms (1, 3, 1024, 1024) -> (256, 256) False False bilinear num_threads=2 0.7X 1.9ms vs 2.9ms (1, 3, 1024, 1024) -> (256, 256) True True bicubic num_threads=2 1.6X 12.3ms vs 7.5ms (1, 3, 1024, 1024) -> (256, 256) True False bicubic num_threads=2 1.0X 3.9ms vs 3.9ms (1, 3, 1024, 1024) -> (256, 256) False True bicubic num_threads=2 1.0X 7.0ms vs 7.3ms (1, 3, 1024, 1024) -> (256, 256) False False bicubic num_threads=2 0.9X 3.0ms vs 3.5ms ``` </details> Benchmark code <details> ```py import operator_benchmark as op_bench import torch """Microbenchmarks for interpolate operator.""" class InterpolateBenchmark(op_bench.TorchBenchmarkBase): def init(self, input_size, output_size, channels_last=False, mode='linear', antialias=False, dtype=torch.float): input_image = torch.randint(0, 256, size=input_size, dtype=torch.uint8, device='cpu') if channels_last: input_image = input_image.contiguous(memory_format=torch.channels_last) self.inputs = { "input_image": input_image, "output_size": output_size, "mode": mode, "antialias": antialias, "dtype":dtype, } self.set_module_name("interpolate") def forward(self, input_image, output_size, mode, antialias, dtype): if dtype == torch.float: input_image = input_image.float() out = torch.nn.functional.interpolate(input_image, size=output_size, mode=mode, align_corners=False, antialias=antialias) if dtype == torch.float: out = out.round().clamp(min=0, max=256).to(torch.uint8) def make_config(): sizes = ( ((224, 224), (64, 64)), ((270, 268), (224, 224)), ((256, 256), (1024, 1024)), ) attrs = [] for (HW1, HW2) in sizes: attrs.append([(1, 3, *HW1), HW2]) # 3 channels # attrs.append([(1, 1, *HW1), HW2]) # 1 channel attrs.append([(1, 3, *HW2), HW1]) # 3 channels # attrs.append([(1, 1, *HW2), HW1]) # 1 channel config = op_bench.config_list( attr_names=["input_size", "output_size"], attrs=attrs, cross_product_configs={ 'channels_last': [True, False], 'mode': ["bilinear", "bicubic"], 'antialias': [True, False], # 'dtype': [torch.float, torch.uint8] # 'dtype': [torch.uint8] 'dtype': [torch.float] }, tags=["short"], ) return config config = make_config() op_bench.generate_pt_test(config, InterpolateBenchmark) if __name__ == "__main__": op_bench.benchmark_runner.main() ``` ```py import re import argparse parser = argparse.ArgumentParser() parser.add_argument("f1", nargs="?", default="main") parser.add_argument("f2", nargs="?", default="new") args = parser.parse_args() with open(args.f1) as f: main = f.readlines() with open(args.f2) as f: new = f.readlines() out = [] for main_line, new_line in zip(main, new): # num_threads=1 # TODO: remove if main_line.startswith("num_threads="): num_threads = int(main_line.split("=")[-1]) if main_line.startswith("# Input"): deets = f"{main_line.strip()}, {num_threads=}" if main_line.startswith("Forward"): main_time = float(main_line.split()[-1]) new_time = float(new_line.split()[-1]) ratio = main_time / new_time fmt = ".1f" if ratio < 3 else ".0f" improv = f"{ratio:{fmt}}X" time_fmt = ",.3f" if new_time < 100 else ",.1f" deets = deets.strip().replace("# Input: ", "") deets = deets.replace(": ", "=") deets = deets.replace("input_size=", "") deets = deets.replace(", output_size=", " -> ") deets = deets.replace("dtype=torch.", "") deets = deets.replace("mode=", "") deets = deets.replace("antialias=", "") deets = deets.replace("channels_last=", "") # deets = deets.replace("channels_last=True, ", "") split = deets.split(",") # size = ','.join(split[:-3]) # mode, dtype, threads = split[-3:] # deets = f"{size:<30} {mode:<15} {dtype:<10} {threads:<15}" size = ','.join(split[:-5]) channels_last, mode, antialias, dtype, threads= split[-5:] deets = f"{size:<33} {channels_last:<7} {antialias:<7} {mode:<10} {threads:<15}" l = f"{deets} {improv:<5} {main_time / 1000:{time_fmt}}ms vs {new_time / 1000:{time_fmt}}ms" out.append(l) def key(s): # s = ''.join(s.split()[1:]) # remove "N.nX" part num_threads = (int(re.findall(r"num_threads=(\d+)", s)[0]),) input_shape, output_shape = re.findall("\(.*?\)", s) input_shape = input_shape[1:-1] # remove parenthesis input_HW = tuple(int(x) for x in input_shape.split(",")[-2:]) input_C = (-int(input_shape.split(",")[1]),) output_HW = tuple(int(x) for x in output_shape[1:-1].split(",")) is_downsample = (output_HW[0] < input_HW[0],) if "linear" in s: mode = "linear" elif "nearest-exact" in s: mode = "nearest-exact" else: # assert "nearest" in s mode = "nearest" mode = (mode,) return is_downsample + input_HW + output_HW + num_threads + input_C + mode for i, l in enumerate(sorted(out, key=key)): if i % 8 == 0: print() # if i % 10 == 0 and i % 40 != 0: # print() # if i % 40 == 0: # print("-" * 100) print(l) ``` </details> Pull Request resolved: pytorch#90771 Approved by: https://github.com/peterbell10, https://github.com/ngimel

… (channels last) (pytorch#96848) ## Description - Based on pytorch#96651 - Improved perfs for vectorized **bilinear** interpolate uint8 RGB-case, **channels last** - unified RGB and RGBA processing code such that RGB input is not copied into RGBA - Performances are more close to Pillow-SIMD (labeled as `Pillow (9.0.0.post1)` in the results) - RGBA case perfs are the same after refactoring (see Source link below) - Fixed mem pointer alignment, added more comments (reviews from pytorch#96651) ## Results - `Pillow (9.0.0.post1)` == Pillow-SIMD ``` [-------------------------------------------------------------------------------------------------- Resize -------------------------------------------------------------------------------------------------] | Pillow (9.0.0.post1) | torch (2.1.0a0+gitd6e220c) PR | torch (2.1.0a0+git2b75955) nightly | Speed-up: PR vs nightly 1 threads: -------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------------- 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=True | 38.674 (+-0.323) | 57.591 (+-0.244) | 131.033 (+-1.448) | 2.275 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (32, 32) aa=False | | 39.471 (+-0.166) | 113.911 (+-1.736) | 2.886 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=True | 128.512 (+-1.916) | 161.592 (+-1.242) | 299.679 (+-2.099) | 1.855 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (224, 224) aa=False | | 150.994 (+-1.180) | 285.331 (+-1.919) | 1.890 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=True | 180.045 (+-2.223) | 220.581 (+-1.363) | 431.057 (+-3.536) | 1.954 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (320, 320) aa=False | | 219.391 (+-1.409) | 429.410 (+-3.620) | 1.957 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=True | 113.911 (+-1.024) | 129.457 (+-1.295) | 459.610 (+-13.322) | 3.550 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (32, 32) aa=False | | 59.800 (+-0.199) | 400.015 (+-11.815) | 6.689 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=True | 283.050 (+-2.664) | 339.143 (+-1.209) | 683.555 (+-4.466) | 2.016 (+-0.000) 3 torch.uint8 channels_last bilinear (520, 520) -> (224, 224) aa=False | | 250.601 (+-1.236) | 603.545 (+-2.644) | 2.408 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=True | 186.723 (+-2.213) | 199.960 (+-1.343) | 860.867 (+-21.763) | 4.305 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (32, 32) aa=False | | 79.188 (+-0.261) | 703.019 (+-25.805) | 8.878 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=True | 412.353 (+-4.476) | 462.230 (+-1.983) | 1101.673 (+-49.299) | 2.383 (+-0.000) 3 torch.uint8 channels_last bilinear (712, 712) -> (224, 224) aa=False | | 327.973 (+-1.852) | 941.062 (+-5.549) | 2.869 (+-0.000) 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=True | 61.191 (+-0.926) | 80.795 (+-0.518) | 160.853 (+-1.506) | 1.991 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=True | 134.488 (+-2.129) | 169.147 (+-1.324) | 327.343 (+-2.846) | 1.935 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=True | 1037.045 (+-24.982) | 938.623 (+-9.010) | 2603.360 (+-20.530) | 2.774 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=True | 52.792 (+-0.613) | 73.692 (+-0.264) | 131.829 (+-1.333) | 1.789 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=True | 139.596 (+-1.944) | 173.778 (+-1.039) | 320.063 (+-2.562) | 1.842 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=True | 690.132 (+-10.946) | 772.758 (+-2.864) | 2036.860 (+-36.109) | 2.636 (+-0.000) 3 torch.uint8 channels_last bilinear (64, 64) -> (224, 224) aa=False | | 78.747 (+-0.799) | 158.479 (+-1.702) | 2.013 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (270, 268) aa=False | | 167.046 (+-1.077) | 322.104 (+-2.764) | 1.928 (+-0.000) 3 torch.uint8 channels_last bilinear (256, 256) -> (1024, 1024) aa=False | | 918.967 (+-5.251) | 2611.388 (+-29.917) | 2.842 (+-0.000) 3 torch.uint8 channels_last bilinear (224, 224) -> (64, 64) aa=False | | 55.336 (+-0.251) | 113.869 (+-1.243) | 2.058 (+-0.000) 3 torch.uint8 channels_last bilinear (270, 268) -> (224, 224) aa=False | | 156.505 (+-1.095) | 299.861 (+-2.710) | 1.916 (+-0.000) 3 torch.uint8 channels_last bilinear (1024, 1024) -> (256, 256) aa=False | | 514.344 (+-1.905) | 1776.796 (+-19.660) | 3.454 (+-0.000) ``` Note: There is no perf regression for other case. There some cases (see Source below) with small speed-ups, for the rest it is roughly around 1.0 +/- 0.1 which may be attributed to noisy measurements ... [Source](https://gist.github.com/vfdev-5/1c0778904a07ce40401306548b9525e8#file-20230329-181023-pr_vs_nightly-speedup-md) ## Context - pytorch#90771 Pull Request resolved: pytorch#96848 Approved by: https://github.com/NicolasHug, https://github.com/peterbell10

NicolasHug and others added 22 commits December 6, 2022 12:19

WIP

3a43ae6

Use double for weights, like in PIL-SIMD

eead8be

Add support for batch_size>1 in AVX version

86b13cb

Add support for C <= 4 in AVX version

48fc70c

Bicubic: fixed fallback (working); AVX version not working yet

c49c09d

Merge branch 'master' into interpolate_uint8_images_linear_cpu_suppor…

e5f36a4

…t_dev

Merge branch 'master' into interpolate_uint8_images_linear_cpu_suppor…

58c421f

…t_dev

Fix support for bicubic

6dcc6d1

AVX: Add support contiguous tensor (not channels_last)

8f2d371

AVX: Clean up

364155d

AVX: more clean up

a2b97ae

Horiz pass passing

db8eb81

Unified Weights computation with fallback and AVX

9034d1d

Refactored repeated code into HelperInterpBase::_compute_indices_int1…

479a9a2

…6_weights_aa

WIP: support antialias=False

f31dfa1

Refactored duplicated code in _compute_indices_int16_weights_aa

9e94be7

Support for antialias=False: done

4d3633e

Merge branch 'interpolate_uint8_images_linear_cpu_support_dev_antiali…

5f7ad39

…as_false' into interpolate_uint8_images_linear_cpu_support_dev

Some cleanup

b053a39

Remove hack to check AVX against fallback

feb8383

More cleanup

873644a

Add PILLOW-SIMD license notice

1899330

github-actions Bot added the module: cpu CPU specific problem (e.g., perf, algorithm) label Dec 13, 2022

NicolasHug mentioned this pull request Dec 13, 2022

WIP Add uint8 support for interpolate on channels_last (1, 3, H,W) CPU images, mode=bilinear, antialias=True #87863

Closed

vfdev-5 reviewed Dec 13, 2022

View reviewed changes

Comment thread aten/src/ATen/native/cpu/UpSampleKernel.cpp

vfdev-5 deleted the interpolate_uint8_images_linear_cpu_support_dev branch February 10, 2023 08:08

NicolasHug mentioned this pull request Feb 10, 2023

TODOs before 0.15 release pytorch/vision#7217

Closed

49 tasks

vfdev-5 mentioned this pull request Feb 15, 2023

Resize relies on interpolate's native uint8 handling pytorch/vision#7262

Closed

This was referenced Mar 13, 2023

Improved perfs for vectorized interpolate cpu uint8 RGB-case #96651

Closed

Improved perfs for vectorized bilinear interpolate cpu uint8 RGB-case (channels last) #96848

Closed

NicolasHug mentioned this pull request Apr 4, 2023

Let Resize() handle uint8 images natively for bilinear mode pytorch/vision#7497

Closed

vfdev-5 mentioned this pull request May 4, 2023

Resize V2 relies on interpolate's native uint8 handling pytorch/vision#7557

Merged

abeleinin mentioned this pull request Aug 17, 2024

add nearest support to interpolate tinygrad/tinygrad#6088

Closed

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Add uint8 support for interpolate for CPU images#90771

Add uint8 support for interpolate for CPU images#90771
NicolasHug wants to merge 46 commits intomasterfrom
interpolate_uint8_images_linear_cpu_support_dev

NicolasHug commented Dec 13, 2022 •

edited

Loading

Uh oh!

pytorch-bot Bot commented Dec 13, 2022 •

edited

Loading

Uh oh!

vadimkantorov commented Dec 13, 2022 •

edited

Loading

Uh oh!

vadimkantorov commented Dec 13, 2022

Uh oh!

Uh oh!

vadimkantorov commented Dec 13, 2022 •

edited

Loading

Uh oh!

vfdev-5 commented Dec 13, 2022

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

9 participants

Conversation

NicolasHug commented Dec 13, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

More benchmark details

Uh oh!

pytorch-bot Bot commented Dec 13, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/90771

❗ 1 Active SEVs

✅ No Failures

Uh oh!

vadimkantorov commented Dec 13, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

vadimkantorov commented Dec 13, 2022

Uh oh!

Uh oh!

vadimkantorov commented Dec 13, 2022 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

vfdev-5 commented Dec 13, 2022

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

9 participants

NicolasHug commented Dec 13, 2022 •

edited

Loading

pytorch-bot Bot commented Dec 13, 2022 •

edited

Loading

vadimkantorov commented Dec 13, 2022 •

edited

Loading

vadimkantorov commented Dec 13, 2022 •

edited

Loading