onnxruntime
[CUDA] Plumb GQA workspace KV-length envelope knob into the L1 estimator (reader-only)
#33014

Merged

[CUDA] Plumb GQA workspace KV-length envelope knob into the L1 estimator (reader-only) #33014

chilo-ms merged 2 commits into main from chilo/gqa-workspace-envelope-plumbing

[CUDA] Plumb GQA workspace KV-length envelope knob into the L1 estima…

249e090a

chilo-ms marked this pull request as ready for review 36 days ago

chilo-ms requested a review from

copilot-pull-request-reviewer 36 days ago

copilot-pull-request-reviewer commented on 2026-10-01

titaiwangms commented on 2026-10-02

[CUDA] Validate GQA workspace envelope configuration

97bf527a

titaiwangms commented on 2026-10-05

titaiwangms approved these changes on 2026-10-05

chilo-ms merged f83bf57d into main 30 days ago

chilo-ms deleted the chilo/gqa-workspace-envelope-plumbing branch 30 days ago

tianleiwu commented on 2026-10-16

Reviewers

titaiwangms

tianleiwu

Assignees

No one assigned

Labels

None yet

Milestone

No milestone