Don't reallocate work item buffers every frame. #17684
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
We were calling `clear()` on the work item buffer table, which caused us to deallocate all the CPU side buffers. This patch changes the logic to instead just clear the buffers individually, but leave their backing stores. This has two consequences: 1. To effectively retain work item buffers from frame to frame, we need to key them off `RetainedViewEntity` values and not the render world `Entity`, which is transient. This PR changes those buffers accordingly. 2. We need to clean up work item buffers that belong to views that went away. Amusingly enough, we actually have a system, `delete_old_work_item_buffers`, that tries to do this already, but it wasn't doing anything because the `clear_batched_gpu_instance_buffers` system already handled that. This patch actually makes the `delete_old_work_item_buffers` system useful, by removing the clearing behavior from `clear_batched_gpu_instance_buffers` and instead making `delete_old_work_item_buffers` delete buffers corresponding to nonexistent views. On Bistro, this PR improves the performance of `batch_and_prepare_binned_render_phase` from 61.2 us to 47.8 us, a 28% speedup.
pcwalton
added
A-Rendering
IceSentry
approved these changes
Feb 4, 2025
|
Code looks good, CI is failing on a unneeded import so everything should be good once it's fixed. |
tychedelia
approved these changes
Feb 4, 2025
tychedelia
added
S-Ready-For-Final-Review
superdump
approved these changes
Feb 5, 2025
pcwalton
added a commit
to pcwalton/bevy
that referenced
this pull request
Feb 8, 2025
PR bevyengine#17684 broke occlusion culling because it neglected to set the indirect parameter offsets for the late mesh preprocessing stage if the work item buffers were already set. This PR moves the update of those values to a new function, `init_work_item_buffers` which is unconditionally called for every phase every frame. Note that there's some complexity in order to handle the case in which occlusion culling was enabled on one frame and disabled on the next, or vice versa. This was necessary in order to make the occlusion culling toggle in the `occlusion_culling` example work again.
github-merge-queue bot
pushed a commit
that referenced
this pull request
Feb 9, 2025
PR #17684 broke occlusion culling because it neglected to set the indirect parameter offsets for the late mesh preprocessing stage if the work item buffers were already set. This PR moves the update of those values to a new function, `init_work_item_buffers`, which is unconditionally called for every phase every frame. Note that there's some complexity in order to handle the case in which occlusion culling was enabled on one frame and disabled on the next, or vice versa. This was necessary in order to make the occlusion culling toggle in the `occlusion_culling` example work again.
mrchantey
pushed a commit
to mrchantey/bevy
that referenced
this pull request
Feb 17, 2025
We were calling `clear()` on the work item buffer table, which caused us to deallocate all the CPU side buffers. This patch changes the logic to instead just clear the buffers individually, but leave their backing stores. This has two consequences: 1. To effectively retain work item buffers from frame to frame, we need to key them off `RetainedViewEntity` values and not the render world `Entity`, which is transient. This PR changes those buffers accordingly. 2. We need to clean up work item buffers that belong to views that went away. Amusingly enough, we actually have a system, `delete_old_work_item_buffers`, that tries to do this already, but it wasn't doing anything because the `clear_batched_gpu_instance_buffers` system already handled that. This patch actually makes the `delete_old_work_item_buffers` system useful, by removing the clearing behavior from `clear_batched_gpu_instance_buffers` and instead making `delete_old_work_item_buffers` delete buffers corresponding to nonexistent views. On Bistro, this PR improves the performance of `batch_and_prepare_binned_render_phase` from 61.2 us to 47.8 us, a 28% speedup. 
mrchantey
pushed a commit
to mrchantey/bevy
that referenced
this pull request
Feb 17, 2025
) PR bevyengine#17684 broke occlusion culling because it neglected to set the indirect parameter offsets for the late mesh preprocessing stage if the work item buffers were already set. This PR moves the update of those values to a new function, `init_work_item_buffers`, which is unconditionally called for every phase every frame. Note that there's some complexity in order to handle the case in which occlusion culling was enabled on one frame and disabled on the next, or vice versa. This was necessary in order to make the occlusion culling toggle in the `occlusion_culling` example work again.
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
We were calling
clear()on the work item buffer table, which caused us to deallocate all the CPU side buffers. This patch changes the logic to instead just clear the buffers individually, but leave their backing stores. This has two consequences:To effectively retain work item buffers from frame to frame, we need to key them off
RetainedViewEntityvalues and not the render worldEntity, which is transient. This PR changes those buffers accordingly.We need to clean up work item buffers that belong to views that went away. Amusingly enough, we actually have a system,
delete_old_work_item_buffers, that tries to do this already, but it wasn't doing anything because theclear_batched_gpu_instance_bufferssystem already handled that. This patch actually makes thedelete_old_work_item_bufferssystem useful, by removing the clearing behavior fromclear_batched_gpu_instance_buffersand instead makingdelete_old_work_item_buffersdelete buffers corresponding to nonexistent views.On Bistro, this PR improves the performance of
batch_and_prepare_binned_render_phasefrom 61.2 us to 47.8 us, a 28% speedup.