EngineX-Hygon/sglang
5
0
Fork 0
Code Issues Pull Requests Actions 7 Projects Releases Wiki Activity

[BUG] fix moe benchmark when bs*seq is small (#3382)

Browse Source
This commit is contained in:
yiakwy-xpu-ml-framework-team
2025-02-08 15:39:44 +08:00
committed by GitHub
parent 4530136e61
commit 64480df495
1 changed files with 2 additions and 2 deletions
Download Patch File Download Diff File Expand all files Collapse all files

4
benchmark/kernels/fused_moe_triton/benchmark_deepseekv3_moe_align_blocks.py
View File

@@ -157,7 +157,7 @@ def calculate_diff(batch_size, seq_len):
) )
sorted_ids_cuda.fill_(topk_ids.numel()) sorted_ids_cuda.fill_(topk_ids.numel())
max_num_m_blocks = max_num_tokens_padded // block_size max_num_m_blocks = max_num_tokens_padded // block_size
expert_ids_cuda = torch.empty( expert_ids_cuda = torch.zeros(
(max_num_m_blocks,), dtype=torch.int32, device=topk_ids.device (max_num_m_blocks,), dtype=torch.int32, device=topk_ids.device
) )
num_tokens_post_pad_cuda = torch.empty( num_tokens_post_pad_cuda = torch.empty(
@@ -172,7 +172,7 @@ def calculate_diff(batch_size, seq_len):
sorted_ids_triton = torch.empty_like(sorted_ids_cuda) sorted_ids_triton = torch.empty_like(sorted_ids_cuda)
sorted_ids_triton.fill_(topk_ids.numel()) sorted_ids_triton.fill_(topk_ids.numel())
expert_ids_triton = torch.empty_like(expert_ids_cuda) expert_ids_triton = torch.zeros_like(expert_ids_cuda)
num_tokens_post_pad_triton = torch.empty_like(num_tokens_post_pad_cuda) num_tokens_post_pad_triton = torch.empty_like(num_tokens_post_pad_cuda)
# compare the performance of cuda and triton implementation # compare the performance of cuda and triton implementation