llama.cpp
[CUDA] : Add SKU-aware FP16/BF16 accumulation precision gating for NVIDIA GPUs
#30672
Merged

[CUDA] : Add SKU-aware FP16/BF16 accumulation precision gating for NVIDIA GPUs #30672

mmalpe-nv
mmalpe-nv cuda: select cuBLAS precision by GPU SKU
df576a6c
github-actions github-actions added ggml
github-actions github-actions added CUDA
mmalpe-nv mmalpe-nv changed the title cuda: select cuBLAS precision by GPU SKU [CUDA]: select cuBLAS precision by GPU SKU 33 days ago
mmalpe-nv mmalpe-nv changed the title [CUDA]: select cuBLAS precision by GPU SKU [CUDA] : Add SKU-aware FP16/BF16 accumulation precision gating for NVIDIA GPUs 33 days ago
ggml-gh-bot
mmalpe-nv mmalpe-nv marked this pull request as ready for review 28 days ago
mmalpe-nv mmalpe-nv requested a review 28 days ago
ynankani
ynankani
mmalpe-nv cuda: fix remaining GB10 macro reference
600c5b53
mmalpe-nv
mmalpe-nv mmalpe-nv closed this 28 days ago
mmalpe-nv mmalpe-nv reopened this 28 days ago
ynankani
ynankani commented on 2026-10-21
mmalpe-nv
mmalpe-nv cuda: review comment addressed, store accumulation preference in devi…
036d8e7e
JohannesGaessler
JohannesGaessler approved these changes on 2026-10-23
IMbackK
ynankani
mmalpe-nv Apply batched suggestions from code review
1aaad3d2
mmalpe-nv Merge branch 'master' of https://github.com/ggml-org/llama.cpp into m…
055b33a5
mmalpe-nv
ORippler
mmalpe-nv addressed Oliver's comment regarding initialize NVML once and install…
cf256c2d
mmalpe-nv mmalpe-nv requested a review 16 hours ago
github-actions github-actions added devops
ORippler
mmalpe-nv merge master and update GB10 references and NVML setup
3d53d4a3
ORippler
ORippler commented on 2026-11-18
mmalpe-nv move precision selection helper into init
ca834154
ORippler
ORippler approved these changes on 2026-11-18
ORippler ORippler merged 2cbfc3cf into master 8 hours ago

Login to write a write a comment.

Login via GitHub

Assignees
No one assigned
Labels
Milestone