[Kernel] Migrate DSA + DSV4 attention kernels to sglang.kernels (RFC #29630, Phase 2.5, 5/7) (#30792)

Co-authored-by: Claude Fable 5 <noreply@anthropic.com>
This commit is contained in:
Xiaoyu Zhang
2026-07-15 11:11:22 +08:00
committed by GitHub
co-authored by Claude Fable 5
parent 4ae9cc3c81
commit ba5be86d42
60 changed files with 663 additions and 583 deletions
@@ -9,8 +9,8 @@ from typing import Tuple
import pytest
import torch
from sglang.srt.layers.attention.dsa.tilelang_kernel import act_quant
from sglang.srt.layers.attention.dsa.triton_kernel import act_quant as act_quant_triton
from sglang.kernels.ops.attention.dsa.tilelang_kernel import act_quant
from sglang.kernels.ops.attention.dsa.triton_kernel import act_quant as act_quant_triton
def benchmark_kernel(
@@ -1,6 +1,6 @@
import torch
from sglang.srt.layers.attention.dsa.index_buf_accessor import (
from sglang.kernels.ops.attention.dsa.index_buf_accessor import (
_get_k_and_s_triton_kernel,
)
@@ -13,7 +13,7 @@ Test coverage:
import pytest
import torch
from sglang.srt.layers.attention.dsa.index_buf_accessor import GetK, GetKAndS, GetS
from sglang.kernels.ops.attention.dsa.index_buf_accessor import GetK, GetKAndS, GetS
class MockDSATokenToKVPool: