Enable fpA_intB GEMM in CUDA builds and add configurable options#29622

Open

tianleiwu wants to merge 18 commits into

microsoft:mainfrom

tianleiwu:tlwu/20260708/fpa_intb_inmem_autotune

Commits on Jul 8, 2026

[CUDA] In-memory tactic autotuning for fpA_intB MatMulNBits
tianleiwu
committed
Enable onnxruntime_USE_FPA_INTB_GEMM by default when CUDA is enabled
tianleiwu
committed
Add ep.cuda.fpa_intb_gemm / fpa_intb_profile_m session-config keys
tianleiwu
committed
Test ep.cuda.fpa_intb_gemm / fpa_intb_profile_m session-config keys
tianleiwu
committed
Merge remote-tracking branch 'origin/main' into tlwu/20260708/fpa_intb_inmem_autotune
tianleiwu
committed
Simplify fpA_intB enable option to boolean; prepacked forces the path
tianleiwu
committed
Merge 'origin/main'
tianleiwu
committed

Commits on Jul 9, 2026

Commits on Jul 13, 2026