Skip to content

Reduce unnecessary kernels and copies in the NSA indexer #22232

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
HaiShaw merged 5 commits into from
Apr 7, 2026
+13 −5
Merged

Reduce unnecessary kernels and copies in the NSA indexer #22232

Changes from all commits
Commits
Show all changes
5 commits
Select commit Hold shift + click to select a range
4588421
Remove redundant clone when using in-place rope kernel
1am9trash Apr 2, 2026
024194f
Use torch compile in indexer with hip
1am9trash Apr 2, 2026
9cbc7f6
Add comment for _update_rope_guarded
1am9trash Apr 2, 2026
f09c845
Merge branch 'sgl-project:main' into optimize_nsa_indexer
1am9trash Apr 7, 2026
82d4150
Merge branch 'main' into optimize_nsa_indexer
HaiShaw Apr 7, 2026
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 13 additions & 5 deletions python/sglang/srt/layers/attention/nsa/nsa_indexer.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -257,13 +257,13 @@ def _weights_proj_bf16_in_fp32_out(self, x: torch.Tensor) -> torch.Tensor:
weights, _ = self.weights_proj(x)
return weights.float()

@torch.compile(dynamic=True) if not _is_hip else lambda f: f
@torch.compile(dynamic=True)
def _project_and_scale_head_gates(self, x: torch.Tensor):
weights = self._weights_proj_bf16_in_fp32_out(x)
weights = weights * self.n_heads**-0.5
return weights

@torch.compile(dynamic=True) if not _is_hip else lambda f: f
@torch.compile(dynamic=True)
def _get_logits_head_gate(self, x: torch.Tensor, q_scale: torch.Tensor):
weights = self._weights_proj_bf16_in_fp32_out(x)
weights = weights * self.n_heads**-0.5
Expand Down Expand Up @@ -318,8 +318,8 @@ def _get_q_k_bf16(

q_rope, k_rope = self.rotary_emb(positions, q_rope, k_rope)

query[..., : self.rope_head_dim] = q_rope.clone()
key[..., : self.rope_head_dim] = k_rope.clone()
self._update_rope_guarded(query[..., : self.rope_head_dim], q_rope)
self._update_rope_guarded(key[..., : self.rope_head_dim], k_rope)

if enable_dual_stream:
current_stream = torch.cuda.current_stream()
Expand Down Expand Up @@ -376,11 +376,19 @@ def _get_k_bf16(
)

_, k_rope = self.rotary_emb(positions, k_rope, k_rope)
key[..., : self.rope_head_dim] = k_rope.clone()
self._update_rope_guarded(key[..., : self.rope_head_dim], k_rope)
key = rotate_activation(key)

return key

@staticmethod
def _update_rope_guarded(dst: torch.Tensor, src: torch.Tensor) -> None:
# On AMD with in-place RoPE kernels, self-aliasing can occur;
# skip write-back when src/dst tensors point to a single memory.
if src.data_ptr() == dst.data_ptr():
return
dst.copy_(src)

def _get_topk_paged(
self,
forward_batch: ForwardBatch,
Expand Down
Loading