Color management: multi-thread IMB_colormanagement_transform_from_byte_threaded simple case #115628

Aras Pranckevicius · 2023-11-30T20:27:37+01:00

2023-11-30 20:27:37 +01:00

This function is only used in Sequencer code to convert source byte images into float images. For a "simple case" where there is no color space conversion needed, it was doing that on a single thread, in two passes over the image (first byte -> float conversion, then alpha premultiply).

Originally this simple code path was added in combination of ce629c5dd9 and 06ac6ded66 (in 2017).

VSE 4K resolution playback, with two image strips where one of them has "convert to float" option, on Windows / Ryzen 5950X:

Overall playback FPS: 3.40 -> 4.03. Still very slow, and a lot of time is spent in allocating, clearing and deleting various temp images, as well as final color conversion for display.
seq_imbuf_to_sequencer_space part: 217.9ms -> 86.6ms. Most of remaining cost is just in just memory allocation+clear, which is why it's not drastically faster due to threading.

This function is only used in Sequencer code to convert source byte images into float images. For a "simple case" where there is no color space conversion needed, it was doing that on a single thread, in two passes over the image (first byte -> float conversion, then alpha premultiply). Originally this simple code path was added in combination of ce629c5dd9 and 06ac6ded66 (in 2017). VSE 4K resolution playback, with two image strips where one of them has "convert to float" option, on Windows / Ryzen 5950X: - Overall playback FPS: 3.40 -> 4.03. Still very slow, and a lot of time is spent in allocating, clearing and deleting various temp images, as well as final color conversion for display. - `seq_imbuf_to_sequencer_space` part: 217.9ms -> 86.6ms. Most of remaining cost is just in just memory allocation+clear, which is why it's not drastically faster due to threading.

❤️ 6

Aras Pranckevicius added 1 commit 2023-11-30 20:27:47 +01:00

buildbot/vexp-code-patch-coordinator Build done. Details

521fce4b83 Color management: multi-thread IMB_colormanagement_transform_from_byte_threaded simple case

This function is only used in Sequencer code to convert source byte
images into float images. For a "simple case" where there is no color
space conversion needed, it was doing that on a single thread, in two
passes over the image (first byte -> float conversion, then alpha
premultiply).

Originally this simple code path was added in combination of ce629c5dd9
and 06ac6ded66 (in 2017).

VSE 4K resolution playback, with two image strips where one of them
has "convert to float" option, on Windows / Ryzen 5950X:
- Overall playback FPS: 3.40 -> 4.03. Still very slow, and a lot of time
  is spent in allocating, clearing and deleting various temp
  images, as well as final color conversion for display.
- seq_imbuf_to_sequencer_space part: 217.9ms -> 86.6ms. Most of remaining
  cost is just in just memory allocation+clear, which is why it's not
  drastically faster due to threading.

Aras Pranckevicius commented

2023-11-30 20:28:24 +01:00

@blender-bot build

Aras Pranckevicius added this to the Video Sequencer project 2023-11-30 20:28:45 +01:00

Aras Pranckevicius requested review from Sergey Sharybin 2023-12-01 05:59:47 +01:00

Sergey Sharybin approved these changes 2023-12-01 15:29:16 +01:00

Aras Pranckevicius merged commit f3ce0645e4 into main

2023-12-01 16:18:57 +01:00

Aras Pranckevicius referenced this issue from a commit

2023-12-01 16:18:58 +01:00

Color management: multi-thread IMB_colormanagement_transform_from_byte_threaded simple case

Aras Pranckevicius deleted branch cm-transform_from_byte-opt

2023-12-01 16:19:00 +01:00

Sign in to join this conversation.

No reviewers

No Label

Download

What's New

Blender Studio

Manual

Developers Blog

Documentation

Benchmark

Blender Conference

Development Fund

One-time Donations

Color management: multi-thread IMB_colormanagement_transform_from_byte_threaded simple case #115628