Skip to content

CUDA: Remove unneded bias/gate dims in fused mmvq#16858

Merged
am17an merged 3 commits intoggml-org:masterfrom
ORippler:osimons/remove_unneded_bias_gate_dims
Nov 1, 2025
Merged

CUDA: Remove unneded bias/gate dims in fused mmvq#16858
am17an merged 3 commits intoggml-org:masterfrom
ORippler:osimons/remove_unneded_bias_gate_dims

Conversation

@ORippler
Copy link
Contributor

Pointed out
here that only a single value is needed per target col per thread

Pointed out
[here](ggml-org#16847 (comment))
that only a single value is needed per target col per thread
Copy link
Collaborator

@JohannesGaessler JohannesGaessler left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Since you're already making another PR, please explicitly unroll the loops over ncols_dst.

@github-actions github-actions bot added Nvidia GPU Issues specific to Nvidia GPUs ggml changes relating to the ggml tensor library for machine learning labels Oct 30, 2025
@ORippler ORippler force-pushed the osimons/remove_unneded_bias_gate_dims branch from c66a33d to f79bd92 Compare October 31, 2025 10:10
@am17an am17an merged commit d3dc9dd into ggml-org:master Nov 1, 2025
71 of 72 checks passed
Anico2 added a commit to Anico2/llama.cpp that referenced this pull request Jan 15, 2026
* CUDA: Remove unneded bias/gate dims in fused mmvq

Pointed out
[here](ggml-org#16847 (comment))
that only a single value is needed per target col per thread

* Apply suggestions from code review

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>

* Fix "Error 991-D: extra braces are nonstandard" during compilation

---------

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
blime4 referenced this pull request in blime4/llama.cpp Feb 5, 2026
* CUDA: Remove unneded bias/gate dims in fused mmvq

Pointed out
[here](ggml-org/llama.cpp#16847 (comment))
that only a single value is needed per target col per thread

* Apply suggestions from code review

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>

* Fix "Error 991-D: extra braces are nonstandard" during compilation

---------

Co-authored-by: Johannes Gäßler <johannesg@5d6.de>
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

ggml changes relating to the ggml tensor library for machine learning Nvidia GPU Issues specific to Nvidia GPUs

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants