biondizzle/vllm
1
0
Fork 0
Code Issues Pull Requests Actions 2 Packages Projects Releases Wiki Activity

[MoE Refactor] Integrate Naive Prepare Finalize into MK (#32567)

Browse Source 
Signed-off-by: Robert Shaw <robshaw@redhat.com>
Signed-off-by: Amir Klein <203507526+amirkl94@users.noreply.github.com>
Co-authored-by: Robert Shaw <robshaw@redhat.com>
Co-authored-by: amirkl94 <203507526+amirkl94@users.noreply.github.com>
This commit is contained in:
Robert Shaw
2026-01-26 20:28:02 -05:00
committed by GitHub
parent 6d86fde09c
commit 5a93b9162b
46 changed files with 1018 additions and 876 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
benchmarks/kernels/benchmark_cutlass_moe_nvfp4.py
View File

@@ -197,7 +197,7 @@ def bench_run(
)
kernel = mk.FusedMoEModularKernel(
MoEPrepareAndFinalizeNoEP(defer_input_quant=True),
MoEPrepareAndFinalizeNoEP(),
CutlassExpertsFp4(
make_dummy_moe_config(),
quant_config=quant_config,
@@ -242,7 +242,7 @@ def bench_run(
)
kernel = mk.FusedMoEModularKernel(
MoEPrepareAndFinalizeNoEP(defer_input_quant=True),
MoEPrepareAndFinalizeNoEP(),
CutlassExpertsFp4(
make_dummy_moe_config(),
quant_config=quant_config,