[ROCm][DSv4.1][Perf] Run the delayed mHC seams through aiter's fused Triton kernel (#58655)

Signed-off-by: Ahmed <Muhammad.Ahmed@amd.com>
This commit is contained in:
Muhammad Ahmed
2026-09-29 10:19:10 +08:00
committed by GitHub
parent 7230dfea50
commit 0af34418e9
6 changed files with 323 additions and 15 deletions
@@ -13,6 +13,7 @@ __all__ = [
"hc_head_fused_kernel_cuda",
"mhc_pre_aiter",
"mhc_pre_delayed_aiter",
"mhc_fused_post_pre_delayed_rms_norm_aiter",
"mhc_post_aiter",
"mhc_fused_post_pre_aiter",
"hc_head_fused_aiter",
+79
View File
@@ -207,6 +207,79 @@ def _mhc_pre_delayed_aiter_fake(
)
def mhc_fused_post_pre_delayed_rms_norm_aiter(
residual: torch.Tensor,
fn: torch.Tensor,
hc_scale: torch.Tensor,
hc_base: torch.Tensor,
rms_eps: float,
hc_pre_eps: float,
hc_sinkhorn_eps: float,
hc_post_mult_value: float,
sinkhorn_repeat: int,
pre_mix: torch.Tensor | None,
sublayer_out: torch.Tensor | None,
post_layer_mix: torch.Tensor | None,
comb_res_mix: torch.Tensor | None,
norm_weight: torch.Tensor,
norm_eps: float,
residual_out: torch.Tensor | None,
) -> tuple[torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor]:
""":func:`mhc_pre_delayed_aiter` with the RMSNorm of the collapse folded in:
``layer_input`` comes back normalised with ``norm_weight`` / ``norm_eps``."""
from vllm._aiter_ops import rocm_aiter_ops
return rocm_aiter_ops.mhc_fused_post_pre_delayed_rms_norm(
residual,
fn,
hc_scale,
hc_base,
rms_eps,
hc_pre_eps,
hc_sinkhorn_eps,
hc_post_mult_value,
sinkhorn_repeat,
pre_mix,
sublayer_out,
post_layer_mix,
comb_res_mix,
norm_weight,
norm_eps,
residual_out,
)
def _mhc_fused_post_pre_delayed_rms_norm_aiter_fake(
residual: torch.Tensor,
fn: torch.Tensor,
hc_scale: torch.Tensor,
hc_base: torch.Tensor,
rms_eps: float,
hc_pre_eps: float,
hc_sinkhorn_eps: float,
hc_post_mult_value: float,
sinkhorn_repeat: int,
pre_mix: torch.Tensor | None,
sublayer_out: torch.Tensor | None,
post_layer_mix: torch.Tensor | None,
comb_res_mix: torch.Tensor | None,
norm_weight: torch.Tensor,
norm_eps: float,
residual_out: torch.Tensor | None,
) -> tuple[torch.Tensor, torch.Tensor, torch.Tensor, torch.Tensor]:
return _mhc_pre_delayed_aiter_fake(
residual,
fn,
hc_scale,
hc_base,
rms_eps,
hc_pre_eps,
hc_sinkhorn_eps,
hc_post_mult_value,
sinkhorn_repeat,
)
def mhc_post_aiter(
x: torch.Tensor,
residual: torch.Tensor,
@@ -335,6 +408,12 @@ direct_register_custom_op(
mutates_args=["residual_out"],
fake_impl=_mhc_pre_delayed_aiter_fake,
)
direct_register_custom_op(
op_name="mhc_fused_post_pre_delayed_rms_norm_aiter",
op_func=mhc_fused_post_pre_delayed_rms_norm_aiter,
mutates_args=["residual_out"],
fake_impl=_mhc_fused_post_pre_delayed_rms_norm_aiter_fake,
)
direct_register_custom_op(
op_name="mhc_post_aiter",
op_func=mhc_post_aiter,