DeepSeek/DeepEP
1
0
Fork 0
mirror of https://github.com/deepseek-ai/DeepEP synced 2025-06-26 18:28:11 +00:00
Code Issues Actions Packages Projects Releases Wiki Activity

Remove all raw tensors for better P2P overlapping

Browse Source
This commit is contained in:
Chenggang Zhao
2025-03-03 14:25:22 +08:00
parent f60306409a
commit 6cc3497df8
4 changed files with 18 additions and 14 deletions
Download Patch File Download Diff File Expand all files Collapse all files

1
csrc/kernels/api.cuh
View File

@@ -132,6 +132,7 @@ void clean_low_latency_buffer(int* clean_0, int num_clean_int_0,
void dispatch(void* packed_recv_x, float* packed_recv_x_scales,
int* packed_recv_src_info, int64_t* packed_recv_layout_range,
int* packed_recv_count,
void* rdma_recv_x, int* rdma_recv_count, void* rdma_x,
const void* x, const int64_t* topk_idx,
int* next_clean, int num_next_clean_int,