gahow/xserv
1
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

kernels/cuda: paged-attention kernel, dispatch, pinned host memory

Browse Source 
CUDA layer for the paged-KV + swap work:
- csrc: new paged_attention.cu plus updates across attention/gemm/norm/
  activation/embedding/reduce kernels and common.cuh.
- xserv-kernels: new dispatch module and kernel-binding updates.
- xserv-cuda: cudaMallocHost/FreeHost bindings + PinnedBuffer (host swap
  pool backing) and offset-aware D2H/H2D copies used to move KV blocks
  between the GPU pool and pinned host memory.

Co-Authored-By: Claude Opus 4.7 <noreply@anthropic.com>
This commit is contained in:
Gahow Wang
2026-05-28 19:58:36 +08:00
parent 3f1c3d429a
commit 4c3f914459
27 changed files with 581 additions and 32 deletions
Download Patch File Download Diff File Expand all files Collapse all files

8
csrc/embedding/transpose.cu
View File

@@ -1,4 +1,5 @@
#include <cuda_bf16.h>
#include "../common.cuh"
// Transpose between [S, H, D] and [H, S, D] layouts (used for RoPE and attention).
// Also handles [S, H*D] → [H, S, D] (reshape_heads) and reverse (merge_heads).
@@ -169,6 +170,7 @@ void launch_reshape_heads_bf16(const void* in, void* out,
int grid = (total + block - 1) / block;
reshape_heads_bf16<<<grid, block, 0, (cudaStream_t)stream>>>(
(const __nv_bfloat16*)in, (__nv_bfloat16*)out, seq_len, num_heads, head_dim);
CUDA_CHECK_LAST_ERROR();
}
void launch_merge_heads_bf16(const void* in, void* out,
@@ -178,6 +180,7 @@ void launch_merge_heads_bf16(const void* in, void* out,
int grid = (total + block - 1) / block;
merge_heads_bf16<<<grid, block, 0, (cudaStream_t)stream>>>(
(const __nv_bfloat16*)in, (__nv_bfloat16*)out, seq_len, num_heads, head_dim);
CUDA_CHECK_LAST_ERROR();
}
void launch_transpose_hsd_to_shd_bf16(const void* in, void* out,
@@ -187,6 +190,7 @@ void launch_transpose_hsd_to_shd_bf16(const void* in, void* out,
int grid = (total + block - 1) / block;
transpose_hsd_to_shd_bf16<<<grid, block, 0, (cudaStream_t)stream>>>(
(const __nv_bfloat16*)in, (__nv_bfloat16*)out, seq_len, num_heads, head_dim);
CUDA_CHECK_LAST_ERROR();
}
void launch_transpose_shd_to_hsd_bf16(const void* in, void* out,
@@ -196,6 +200,7 @@ void launch_transpose_shd_to_hsd_bf16(const void* in, void* out,
int grid = (total + block - 1) / block;
transpose_shd_to_hsd_bf16<<<grid, block, 0, (cudaStream_t)stream>>>(
(const __nv_bfloat16*)in, (__nv_bfloat16*)out, seq_len, num_heads, head_dim);
CUDA_CHECK_LAST_ERROR();
}
void launch_repeat_kv_bf16(const void* in, void* out,
@@ -205,6 +210,7 @@ void launch_repeat_kv_bf16(const void* in, void* out,
int grid = (total + block - 1) / block;
repeat_kv_bf16<<<grid, block, 0, (cudaStream_t)stream>>>(
(const __nv_bfloat16*)in, (__nv_bfloat16*)out, kv_heads, n_rep, seq_len, head_dim);
CUDA_CHECK_LAST_ERROR();
}
void launch_strided_copy_bf16(const void* in, void* out, int numel, int ndim,
@@ -217,6 +223,7 @@ void launch_strided_copy_bf16(const void* in, void* out, int numel, int ndim,
(const __nv_bfloat16*)in, (__nv_bfloat16*)out, numel, ndim,
shape0, shape1, shape2, shape3,
in_stride0, in_stride1, in_stride2, in_stride3, in_offset);
CUDA_CHECK_LAST_ERROR();
}
void launch_strided_copy_f32(const void* in, void* out, int numel, int ndim,
@@ -229,6 +236,7 @@ void launch_strided_copy_f32(const void* in, void* out, int numel, int ndim,
(const float*)in, (float*)out, numel, ndim,
shape0, shape1, shape2, shape3,
in_stride0, in_stride1, in_stride2, in_stride3, in_offset);
CUDA_CHECK_LAST_ERROR();
}
}