llama.cpp
vulkan: use fp32 in coopmat2 q4_k dequant function
#12309
Merged

vulkan: use fp32 in coopmat2 q4_k dequant function #12309

0cc4m merged 3 commits into ggml-org:master from jeffbolznv:cm2_q4_k_fp32
jeffbolznv
jeffbolznv jeffbolznv requested a review from 0cc4m 0cc4m 190 days ago
github-actions github-actions added Vulkan
github-actions github-actions added ggml
jeffbolznv vulkan: Adjust coopmat2 tile sizes and selection heuristic
1577cfdd
jeffbolznv vulkan: Pad N dimension of B matrix for coopmat2 perf, to avoid bound…
c5f29205
jeffbolznv vulkan: use fp32 in coopmat2 q4_k dequant function
717fd25e
jeffbolznv jeffbolznv force pushed from 458c70ab to 717fd25e 189 days ago
0cc4m
0cc4m approved these changes on 2025-03-17
0cc4m 0cc4m merged f07690c9 into master 183 days ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone