onnxruntime
Upgrade cudnn_frontend to 1.24 and enable cuDNN SDPA for MHA/GQA
#28849
Merged

Upgrade cudnn_frontend to 1.24 and enable cuDNN SDPA for MHA/GQA #28849

tianleiwu
tianleiwu cudnn frontend 1.24 for MHA and GQA
59a34edf
Copilot Fix CUDA plugin Windows link failure by linking NVRTC
9710adfe
tianleiwu tianleiwu requested a review from copilot-pull-request-reviewer copilot-pull-request-reviewer 101 days ago
copilot-pull-request-reviewer
copilot-pull-request-reviewer commented on 2026-06-08
tianleiwu address feedbacks
30d3cb11
tianleiwu tianleiwu marked this pull request as ready for review 100 days ago
tianleiwu tianleiwu requested a review from kunal-vaishnavi kunal-vaishnavi 100 days ago
tianleiwu tianleiwu requested a review from titaiwangms titaiwangms 100 days ago
titaiwangms
kunal-vaishnavi
kunal-vaishnavi commented on 2026-06-08
kunal-vaishnavi
kunal-vaishnavi commented on 2026-06-08
kunal-vaishnavi
kunal-vaishnavi commented on 2026-06-08
kunal-vaishnavi
kunal-vaishnavi dismissed these changes on 2026-06-08
tianleiwu address review feedback: reuse computed members for cuDNN flash flags
0d5ad416
tianleiwu Fix MHA cuDNN SDPA auto path to honor is_stable(); drop stale GQA com…
46e35c9c
tianleiwu
tianleiwu tianleiwu dismissed their stale review via 46e35c9c 100 days ago
tianleiwu
kunal-vaishnavi
kunal-vaishnavi approved these changes on 2026-06-08
titaiwangms
titaiwangms approved these changes on 2026-06-09
tianleiwu tianleiwu merged fc7a9f0e into main 100 days ago
tianleiwu tianleiwu deleted the tlwu/20260607/upgrade_cudnn_frontend branch 100 days ago

Login to write a write a comment.

Login via GitHub

Assignees
No one assigned
Labels
Milestone