[AMD] Skip x_scale.transpose+contiguous before bpreshuffle GEMM via upstream pre-transposed scale#24476

Open

Jacob0226 wants to merge 2 commits intosgl-project:mainfrom

Jacob0226:jacob/fused-transpose-scale

Commits on May 6, 2026

[FP8] aiter w8a8: skip x_scale.transpose+contiguous when upstream pre-transposed
Jacob0226
committed
[FP8] Extend transpose_scale optimization to o_proj (fused_flatten path)
Jacob0226
committed