llama.cpp
vulkan: optimizations for deepseek prompt processing
#14555
Merged

vulkan: optimizations for deepseek prompt processing #14555

0cc4m merged 4 commits into ggml-org:master from jeffbolznv:deepseek_opts2
jeffbolznv
jeffbolznv vulkan: allow unclamped loads in coopmat2 mul_mat_id shader
129a0f17
jeffbolznv vulkan: increase coopmat2 mul_mat_id tile size
b54ddba8
jeffbolznv vulkan: optimize mat_mul_id row_ids search to batch loads, and port t…
2b540868
jeffbolznv vulkan: use smaller FA row size when head size is large. applies to b…
bd8e0bfb
jeffbolznv jeffbolznv requested a review from 0cc4m 0cc4m 96 days ago
github-actions github-actions added Vulkan
github-actions github-actions added ggml
0cc4m
0cc4m approved these changes on 2025-07-12
0cc4m 0cc4m merged 98197e5c into master 91 days ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone