vulkan: split mul_mat into multiple dispatches to avoid overflow#19509
Merged
0cc4m merged 2 commits intoggml-org:masterfrom Feb 18, 2026
Merged
vulkan: split mul_mat into multiple dispatches to avoid overflow#195090cc4m merged 2 commits intoggml-org:masterfrom
0cc4m merged 2 commits intoggml-org:masterfrom
Conversation
github-actions
bot
added
testing
The batch dimensions can be greater than the max workgroup count limit, in which case we need to split into multiple dispatches and pass the base index through a push constant. Fall back for the less common p021 and nc variants.
Contributor
Author
|
The new tests are failing on multiple backends, I'll move them to a separate PR so this isn't blocked. |
jeffbolznv
force-pushed
the
mul_mat_batch_overflow
branch
from
ccb4ed3 to
66d7c14
Compare
0cc4m
reviewed
Feb 13, 2026
ggml/src/ggml-vulkan/ggml-vulkan.cpp
Outdated
| while (base_work_group_z < batch) { | ||
| uint32_t groups_z = std::min(batch - base_work_group_z, ctx->device->properties.limits.maxComputeWorkGroupCount[2]); | ||
|
|
||
| ggml_pipeline_request_descriptor_sets(ctx, pipeline, 1); |
Contributor
There was a problem hiding this comment.
Why request the descriptor sets in the loop and not before? It's not gonna retrigger pipeline compile of course, but will ping the descriptor pools more than necessary.
Contributor
Author
There was a problem hiding this comment.
ok, moved it out of the loop for now. Eventually, I'd like to not have to explicitly call this anywhere.
Contributor
There was a problem hiding this comment.
It should be possible to automatically request one when grabbing the pipeline and to allocate the sets on demand before dispatch.
Contributor
Author
There was a problem hiding this comment.
Yeah, the main catch right now is that some shaders won't get their wg_denoms initialized until after this call.
0cc4m
approved these changes
Feb 18, 2026
liparetejas
pushed a commit
to liparetejas/llama.cpp
that referenced
this pull request
Feb 23, 2026
…l-org#19509) * vulkan: split mul_mat into multiple dispatches to avoid overflow The batch dimensions can be greater than the max workgroup count limit, in which case we need to split into multiple dispatches and pass the base index through a push constant. Fall back for the less common p021 and nc variants. * address feedback
bartowski1182
pushed a commit
to bartowski1182/llama.cpp
that referenced
this pull request
Mar 2, 2026
…l-org#19509) * vulkan: split mul_mat into multiple dispatches to avoid overflow The batch dimensions can be greater than the max workgroup count limit, in which case we need to split into multiple dispatches and pass the base index through a push constant. Fall back for the less common p021 and nc variants. * address feedback
ArberSephirotheca
pushed a commit
to ArberSephirotheca/llama.cpp
that referenced
this pull request
Mar 3, 2026
…l-org#19509) * vulkan: split mul_mat into multiple dispatches to avoid overflow The batch dimensions can be greater than the max workgroup count limit, in which case we need to split into multiple dispatches and pass the base index through a push constant. Fall back for the less common p021 and nc variants. * address feedback
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
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
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.
The batch dimensions can be greater than the max workgroup count limit, in which case we need to split into multiple dispatches and pass the base index through a push constant.
Fall back for the less common p021 and nc variants.
Fixes #19471.