Zero init textures by Wumpf · Pull Request #1688 · gfx-rs/wgpu

Wumpf · 2021-07-18T22:03:22Z

Connections
Fixes #1102
Some preparatory work happened in: #2081, #2046, #1905, #1839

Description
Implements zero texture init using some of the same building blocks & patterns already used for buffer zero init.

Moved "inter-submit" init code to a new mod which also has utlitities to tackle the rare texure inits ("fixups") that may happen due to a discarded texture being used within an encoder/cmd-buf. The machinery for this is a bit clunky right now and may be improved down the line.
Special-special handling is given for depth/stencil discards, there are comment blocks that should explain the situation and approach (if those aren't clear enough this needs fixing as well).

Testing
Started out with new / rearranged player tests. For testing the more complicated sideffects of render target discards, a new regular unit test is added.
Overall, I believe testing is about the bare minimum that we can run with for now, but devising good ones is challenging on its own as so often.

player/tests/data/all.ron

sadly, they are typically passing even if texture zero init isn't doing its job However, they form nice isolated examples for testing out texture initialization It could be possible to dirty texture memory prior to ensure zero init did the job

…rtarget

* queue submit * write_texture

enforce COPY_DST usage now on all textures

init_tracker has now a `discard` function to get single data points back to uninitialized use new standardized partition_point function

missing: * init on the fly if discarded is found within command buffer * handle discarding only stencil or only depth

…through utility function allows to resolve discard/init_action interactions

CommandBufferTextureMemoryActions is now fully encapsulated

…ing copies

kvark

I've done the first review pass. This is grand piece of work!
When the group was deciding to zero-init resources, I didn't expect this to be associated with so much complexity. And I hoped this would work together with the state tracker, while you've shown that it's better off to be a completely independent system.

wgpu-core/src/command/memory_init.rs

wgpu/tests/zero_init_texture_after_discard.rs

wgpu-core/src/command/memory_init.rs

Wumpf · 2021-10-25T19:10:33Z

I didn't expect this to be associated with so much complexity

neither did I. Very unlikely I would have promised to look after this thing if I knew 😄
how does the old saying go, we don't do these things because they are easy, we do them because we thought they were easy

… hygenie

… init, cleanups

kvark · 2021-10-28T03:24:06Z

We have a bit of clippy complains, otherwise let's ship it!

Wumpf · 2021-10-28T19:28:54Z

🥳 finally, what a relief!

When no work is submitted for a frame, presenting the surface results in a timeout due to no work having been submitted. Fixes gfx-rs#3189. This flag was added in gfx-rs#1892 with a note that it was going to be temporary until gfx-rs#1688 landed.

When no work is submitted for a frame, presenting the surface results in a timeout due to no work having been submitted. Fixes #3189. This flag was added in #1892 with a note that it was going to be temporary until #1688 landed.

kvark reviewed Jul 22, 2021

View reviewed changes

player/tests/data/all.ron Show resolved Hide resolved

Wumpf force-pushed the zero-init-textures branch 2 times, most recently from e4cd399 to dfaffa6 Compare August 21, 2021 15:10

This was referenced Aug 21, 2021

Memory init tracker refactor #1839

Merged

fix buffer collapse in initialize_buffer_memory #1840

Merged

Wumpf force-pushed the zero-init-textures branch from dfaffa6 to bbb9202 Compare August 28, 2021 08:32

kvark mentioned this pull request Sep 1, 2021

More validation of the surface texture lifetime #1892

Merged

Wumpf mentioned this pull request Sep 4, 2021

Reintroduce clear_texture Metal/Vulkan/DX12 #1905

Merged

Wumpf force-pushed the zero-init-textures branch 2 times, most recently from 533a2f8 to ae11fe2 Compare September 12, 2021 20:54

Wumpf mentioned this pull request Oct 9, 2021

Move clear_texture impl from wgpu-hal to wgpu-core #2046

Merged

Wumpf mentioned this pull request Oct 16, 2021

init_tracker has now a discard function to get single data points back to uninitialized #2081

Merged

Wumpf added 17 commits October 24, 2021 17:24

Added tests for texture zero init

26b38ce

sadly, they are typically passing even if texture zero init isn't doing its job However, they form nice isolated examples for testing out texture initialization It could be possible to dirty texture memory prior to ensure zero init did the job

texture init tracker

93baf5d

tracking texture init requirements for bind group, transfer and rende…

c67c644

…rtarget

texture clears for texture init

f10d7a5

* queue submit * write_texture

Enforce presence of either render target or copy_dst flag

a8412b9

clear render targets also with using buffer copies

95d54e5

enforce COPY_DST usage now on all textures

adjust ImageSubresourceRange.layer_range calculation for 3D textures

23bc55c

init_tracker has now a `discard` function to get single data points back to uninitialized use new standardized partition_point function

track init state for discarded textures from renderpasses

22b9e89

missing: * init on the fly if discarded is found within command buffer * handle discarding only stencil or only depth

added tests for zero init after discard

d0c949c

tracking discarded surfaces now in separate struct, piping all inits …

e1924db

…through utility function allows to resolve discard/init_action interactions

Move various memory init code to separate mod

0a73ede

CommandBufferTextureMemoryActions is now fully encapsulated

implemented discard init fixups for everything but renderpass

50c2732

render passes also cause now discard fixups

cd023bf

fixup_discarded_surfaces takes now an iterator instead of Drain

35116ec

Add memory init test for discarding depth targets

2a25ccc

handle divergently discarded depth/stencil target

e63a90f

comment & clippy fixes

bb0aaab

Wumpf force-pushed the zero-init-textures branch from ae11fe2 to bb0aaab Compare October 24, 2021 16:56

fix collect_zero_buffer_copies_for_clear_texture yielding block break…

87dc9b0

…ing copies

Wumpf changed the title ~~[draft] Zero init textures~~ Zero init textures Oct 24, 2021

Wumpf marked this pull request as ready for review October 24, 2021 17:31

kvark reviewed Oct 25, 2021

View reviewed changes

Wumpf added 2 commits October 25, 2021 21:31

[pr feedback] minor cleanup in zero_init_texture_after_discard, use…

a79fe0e

… hygenie

[pr feedback] fix bug in ImageSubresourceRange range utils

b48d444

Wumpf force-pushed the zero-init-textures branch from 3047510 to b48d444 Compare October 25, 2021 19:32

[pr feedback] fix texture tracker check, bundle transition_texture on…

50778a5

… init, cleanups

kvark assigned Wumpf Oct 27, 2021

Wumpf added 2 commits October 27, 2021 22:49

Implemented drop for InitTrackerDrain

b9c63aa

remove incorrect comment about extents in add_pass_texture_init_actions

65c4209

kvark approved these changes Oct 28, 2021

View reviewed changes

Fix unit test & clippy issues in init_tracker

a7613e7

kvark approved these changes Oct 28, 2021

View reviewed changes

kvark merged commit 9887169 into gfx-rs:master Oct 28, 2021

Wumpf deleted the zero-init-textures branch October 28, 2021 19:28

sandydoo mentioned this pull request Oct 30, 2021

Metal backend not working because of missing surface capabilities #2135

Closed

ElectronicRU mentioned this pull request Jan 2, 2022

Don't create array layer trackers for 3D textures. #2348

Merged

teoxoy mentioned this pull request Feb 24, 2023

"No work has been submitted for this frame" timeout #3189

Closed

waywardmonkeys mentioned this pull request Feb 5, 2024

Remove TextureInner::Surface::has_work. #5200

Merged

6 tasks

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Zero init textures#1688

Zero init textures#1688
kvark merged 24 commits intogfx-rs:masterfrom
Wumpf:zero-init-textures

Wumpf commented Jul 18, 2021 •

edited

Loading

Uh oh!

Uh oh!

kvark left a comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Wumpf commented Oct 25, 2021

Uh oh!

kvark commented Oct 28, 2021

Uh oh!

Wumpf commented Oct 28, 2021

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Conversation

Wumpf commented Jul 18, 2021 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

Uh oh!

kvark left a comment

Choose a reason for hiding this comment

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Uh oh!

Wumpf commented Oct 25, 2021

Uh oh!

kvark commented Oct 28, 2021

Uh oh!

Wumpf commented Oct 28, 2021

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

2 participants

Wumpf commented Jul 18, 2021 •

edited

Loading