llama.cpp
vulkan: subgroup size tuning
#12087
Merged

vulkan: subgroup size tuning #12087

daniandtheweb
github-actions github-actions added Vulkan
github-actions github-actions added ggml
0cc4m
0cc4m
daniandtheweb daniandtheweb force pushed to 7037e948 317 days ago
daniandtheweb daniandtheweb force pushed 316 days ago
daniandtheweb daniandtheweb force pushed 316 days ago
0cc4m
daniandtheweb
daniandtheweb vulkan: subgroup size test
53b69a31
0cc4m Vulkan: Add device architecture enum and logic to recognize AMD gener…
85e15e6b
daniandtheweb vulkan: use new architecture logic to specify subgroup size
76955414
daniandtheweb daniandtheweb force pushed to 76955414 314 days ago
daniandtheweb daniandtheweb marked this pull request as ready for review 314 days ago
daniandtheweb daniandtheweb changed the title vulkan: subgroup size test vulkan: subgroup size tuning 314 days ago
daniandtheweb
0cc4m
daniandtheweb
0cc4m
daniandtheweb Initial vulkan subgroup size tuning for RDNA3
43c3e6fd
daniandtheweb
0cc4m
daniandtheweb vulkan: commonize RDNA subgroup tuning
c41619d0
daniandtheweb
0cc4m
0cc4m approved these changes on 2025-03-12
0cc4m
0cc4m commented on 2025-03-12
daniandtheweb vulkan: override subgroup size if required_subgroup_size = 0
1c17520f
daniandtheweb vulkan: disable warp 32 for RDNA3
afb5c2dc
daniandtheweb
0cc4m
daniandtheweb
daniandtheweb
daniandtheweb
daniandtheweb
daniandtheweb vulkan: fine tuned RDNA1 subgroup sizes
29e81049
0cc4m
0cc4m
0cc4m commented on 2025-03-16
0cc4m 0cc4m requested a review from 0cc4m 0cc4m 309 days ago
0cc4m
daniandtheweb
0cc4m
0cc4m
daniandtheweb vulkan: adjusted subgroup size map
bf7352e1
daniandtheweb vulkan: fixed RDNA2 subgroup map
d43537ad
daniandtheweb
0cc4m
0cc4m approved these changes on 2025-03-17
0cc4m 0cc4m merged cf2270e4 into master 308 days ago

Login to write a write a comment.

Login via GitHub

Reviewers
Assignees
No one assigned
Labels
Milestone