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

[Kernel] Integrate CUTLASS MoE kernel with PPLX (#18762)

Browse Source 
Signed-off-by: ElizaWszola <ewszola@redhat.com>
Signed-off-by: Tyler Michael Smith <tyler@neuralmagic.com>
Co-authored-by: Tyler Michael Smith <tyler@neuralmagic.com>
This commit is contained in:
ElizaWszola
2025-06-07 03:26:11 +02:00
committed by GitHub
parent 6e0cd10f72
commit 84166fee97
26 changed files with 918 additions and 409 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
tests/kernels/moe/test_cutlass_moe.py
View File

@@ -193,14 +193,10 @@ def run_8_bit(moe_tensors: MOETensors8Bit,
kwargs = {
'a': moe_tensors.a,
'w1_q': moe_tensors.w1_q.transpose(1, 2), # type: ignore[union-attr]
'w2_q': moe_tensors.w2_q.transpose(1, 2), # type: ignore[union-attr]
'w1_q': moe_tensors.w1_q, # type: ignore[union-attr]
'w2_q': moe_tensors.w2_q, # type: ignore[union-attr]
'topk_weights': topk_weights,
'topk_ids': topk_ids,
'ab_strides1': moe_tensors.ab_strides1,
'c_strides1': moe_tensors.c_strides1,
'ab_strides2': moe_tensors.ab_strides2,
'c_strides2': moe_tensors.c_strides2,
'w1_scale': moe_tensors.w1_scale,
'w2_scale': moe_tensors.w2_scale,
'a1_scale': moe_tensors.a_scale