Revert "[Perf][GLM-5.2] Blackwell decode optimizations" (#49768)
This commit is contained in:
@@ -330,10 +330,6 @@ STABLE_TORCH_LIBRARY_FRAGMENT(_C, ops) {
|
||||
// conditionally compiled so impl registration is in source file
|
||||
ops.def("fp32_router_gemm(Tensor! output, Tensor mat_a, Tensor mat_b) -> ()");
|
||||
|
||||
// BF16 skinny GEMM (M<=32, weight-BW-bound shapes, e.g. MTP eh_proj).
|
||||
// conditionally compiled so impl registration is in source file
|
||||
ops.def("bf16_skinny_gemm(Tensor! output, Tensor mat_a, Tensor mat_b) -> ()");
|
||||
|
||||
// reorder weight for AllSpark Ampere W8A16 Fused Gemm kernel
|
||||
ops.def(
|
||||
"rearrange_kn_weight_as_n32k16_order(Tensor b_qweight, Tensor b_scales, "
|
||||
|
||||
Reference in New Issue
Block a user