EngineX/xc-llm-ascend
3
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
Files
42bcad7e9b68cf92d9b98a9072eff10edc8d6f33
xc-llm-ascend/vllm_ascend/ops/fused_moe/moe_mlp.py
chenxi-hh 42bcad7e9b GMM custom operator optimization in small batch scenarios (#7100) 
### What this PR does / why we need it?
GMM custom operator optimization in small batch scenarios

### How was this patch tested?

Qwen3-30B input: 4k, output: 1k

batch 1：
TPOT 7.9 ms -> 7.0 ms
Output Token Throughput 125.4651 token/s -> 140.6278 token/s

batch 2：
TPOT 9.4 ms -> 8.8 ms
Output Token Throughput 211.8187 token/s -> 225.2254 token/s

batch 16：
TPOT 13.6 ms -> 13.5 ms
Output Token Throughput 1159.8213 token/s -> 1165.0982 token/s

- vLLM version: v0.16.0
- vLLM main:
4034c3d32e

---------

Signed-off-by: chenxi-hh <chen464822955@163.com>
2026-03-19 16:10:30 +08:00

467 lines
18 KiB
Python
Raw Blame History

# Copyright (c) 2025 Huawei Technologies Co., Ltd. All Rights Reserved.
# Copyright 2023 The vLLM team.
#
# Licensed under the Apache License, Version 2.0 (the "License");
# you may not use this file except in compliance with the License.
# You may obtain a copy of the License at
#
# http://www.apache.org/licenses/LICENSE-2.0
#
# Unless required by applicable law or agreed to in writing, software
# distributed under the License is distributed on an "AS IS" BASIS,
# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
# See the License for the specific language governing permissions and
# limitations under the License.
# This file is a part of the vllm-ascend project.
import torch
import torch_npu
from torch.nn.functional import pad
from vllm.forward_context import get_forward_context
from vllm.triton_utils import HAS_TRITON
from vllm_ascend.ascend_forward_context import _EXTRA_CTX, MoECommType
from vllm_ascend.device.device_op import DeviceOperator
from vllm_ascend.device.mxfp_compat import (
ensure_mxfp8_moe_available,
)
from vllm_ascend.ops.activation import AscendSwigluOAIAndMul
from vllm_ascend.utils import (
dispose_tensor,
enable_custom_op,
get_weight_prefetch_method,
)
def _custom_gmm_swiglu_enabled(fusion, dynamic_eplb):
return fusion and dynamic_eplb and enable_custom_op()
def cumsum_group_list(
group_list: torch.Tensor, src_list_type: int, dst_list_type: int, active_num: int = 0, expert_num: int = 0
) -> torch.Tensor:
if src_list_type not in [0, 1, 2]:
raise ValueError(f"group_list_type should be in [0, 1, 2], but received {src_list_type}")
if src_list_type == dst_list_type:
return group_list
if src_list_type == 1 and dst_list_type == 0:
return group_list.cumsum(dim=0)
if src_list_type == 0 and dst_list_type == 1:
group_diff = torch.diff(group_list)
new_group = torch.cat([group_list[0].unsqueeze(0), group_diff], dim=0)
return new_group
if src_list_type == 2 and dst_list_type == 0:
experts = pad(group_list[:, 0], (1, 0))
tokens = pad(group_list[:, 1].cumsum(dim=0), (1, 0))
cumsum_group_list = torch.full(
size=(expert_num,), fill_value=active_num, dtype=group_list.dtype, device=group_list.device
)
for i, (start, end) in enumerate(zip(experts[:-1], experts[1:])):
if end > start:
cumsum_group_list[start:end] = tokens[i]
return cumsum_group_list
raise NotImplementedError(
f"Conversion from src_list_type={src_list_type} to dst_list_type={dst_list_type} is not implemented yet. "
"This feature is under development."
)
def _require_single_tensor_for_swiglu_quant(
tensor_or_list: list[torch.Tensor] | torch.Tensor, *, name: str
) -> torch.Tensor:
if isinstance(tensor_or_list, list):
if len(tensor_or_list) != 1:
raise ValueError(f"{name} must be a tensor or a single-element list, but got {len(tensor_or_list)}.")
return tensor_or_list[0]
return tensor_or_list
def quant_apply_mlp(
hidden_states: torch.Tensor,
w1: list[torch.Tensor] | torch.Tensor,
w1_scale: list[torch.Tensor] | torch.Tensor,
w2: list[torch.Tensor] | torch.Tensor,
w2_scale: list[torch.Tensor] | torch.Tensor,
group_list: torch.Tensor,
group_list_type: int = 1,
dynamic_scale: torch.Tensor = None,
w1_scale_bias: torch.Tensor = None,
w2_scale_bias: torch.Tensor = None,
w1_offset: torch.Tensor | None = None,
w2_offset: torch.Tensor | None = None,
fusion: bool = False,
dynamic_eplb: bool = False,
**kwargs,
) -> torch.Tensor:
# TODO(linfeng): Current massive parameter passing is quite severe; parameter differences introduced by different
# quantization modes will be consolidated into a dataclass in a follow-up.
use_mxfp_quant = kwargs.get("use_mxfp_quant", False)
act_quant_type = torch.float8_e4m3fn
weight_quant_type = None
scale_type = None
per_token_scale_type = None
use_bf16 = True
input_hidden_dtype = hidden_states.dtype
use_gmm_swiglu_quant_fusion = use_mxfp_quant or (fusion and not dynamic_eplb)
if use_mxfp_quant:
act_quant_type = kwargs.get("act_quant_type", torch.float8_e4m3fn)
weight_quant_type = kwargs.get("weight_quant_type", torch.float8_e4m3fn)
scale_type = kwargs.get("scale_type")
per_token_scale_type = kwargs.get("per_token_scale_type")
use_bf16 = kwargs.get("use_bf16", True)
ensure_mxfp8_moe_available("MXFP MoE MLP path")
if w1_scale_bias is not None or w2_scale_bias is not None:
raise NotImplementedError("MXFP path does not support scale_bias yet.")
if w1_offset is not None or w2_offset is not None:
raise NotImplementedError("MXFP path does not support antiquant offset yet.")
if w1_offset is not None:
unquantized_hidden_states = hidden_states
quantized_hidden_states = None
elif dynamic_scale is None:
unquantized_hidden_states = hidden_states
hidden_states, pertoken_scale = DeviceOperator.npu_dynamic_quant(
hidden_states=hidden_states,
dynamic_scale=None,
act_quant_type=act_quant_type,
use_mxfp_quant=use_mxfp_quant,
)
dispose_tensor(unquantized_hidden_states)
quantized_hidden_states = None
else:
unquantized_hidden_states = None
pertoken_scale = dynamic_scale
quantized_hidden_states = hidden_states
bias1, bias2 = None, None
_output_dtype = w2_scale[0].dtype if isinstance(w2_scale, list) else w2_scale.dtype
weight_prefetch_method = get_weight_prefetch_method()
if weight_prefetch_method:
weight_prefetch_method.maybe_prefetch_moe_weight_postprocess(hidden_states)
is_mc2 = _EXTRA_CTX.moe_comm_type == MoECommType.MC2
if w1_scale_bias is None and w1_offset is None and is_mc2:
if _custom_gmm_swiglu_enabled(fusion, dynamic_eplb) and not use_mxfp_quant:
# gmm1: gate_up_proj & act_fn: swiglu
hidden_states, swiglu_out_scale, _ = torch.ops._C_ascend.grouped_matmul_swiglu_quant_weight_nz_tensor_list(
x=hidden_states,
weight=w1,
weight_scale=w1_scale,
x_scale=pertoken_scale,
group_list=cumsum_group_list(group_list, group_list_type, 0),
)
elif use_gmm_swiglu_quant_fusion:
# gmm1: gate_up_proj & act_fn: swiglu
hidden_states, swiglu_out_scale, _ = DeviceOperator.npu_grouped_matmul_swiglu_quant(
x=hidden_states,
weight=_require_single_tensor_for_swiglu_quant(w1, name="w1"),
group_list=cumsum_group_list(group_list, group_list_type, 0),
weight_scale=_require_single_tensor_for_swiglu_quant(w1_scale, name="w1_scale"),
x_scale=pertoken_scale,
bias=None,
use_mxfp_quant=use_mxfp_quant,
)
if quantized_hidden_states is not None:
dispose_tensor(quantized_hidden_states)
else:
if w1_scale[0].dtype != torch.float32:
w1_scale[0] = w1_scale[0].to(torch.float32)
# gmm1: gate_up_proj
hidden_states = torch_npu.npu_grouped_matmul(
x=[hidden_states],
weight=w1,
split_item=3,
group_list_type=group_list_type,
group_type=0,
group_list=group_list,
output_dtype=torch.int32,
)[0]
if quantized_hidden_states is not None:
dispose_tensor(quantized_hidden_states)
# act_fn: swiglu
hidden_states, swiglu_out_scale = torch_npu.npu_dequant_swiglu_quant(
x=hidden_states,
weight_scale=w1_scale[0],
activation_scale=pertoken_scale,
bias=None,
quant_scale=None,
quant_offset=None,
group_index=cumsum_group_list(group_list, group_list_type, 1),
activate_left=True,
quant_mode=1,
)
# gmm2: down_proj
hidden_states = DeviceOperator.npu_grouped_matmul_gmm2(
hidden_states=hidden_states,
weight=w2,
weight_scale=w2_scale,
per_token_scale=swiglu_out_scale,
group_list=group_list,
group_list_type=group_list_type,
input_dtype=input_hidden_dtype,
act_quant_type=act_quant_type,
weight_quant_type=weight_quant_type,
scale_type=scale_type,
per_token_scale_type=per_token_scale_type,
use_bf16=use_bf16,
use_mxfp_quant=use_mxfp_quant,
bias=None,
fallback_output_dtype=w2_scale[0].dtype if isinstance(w2_scale, list) else w2_scale.dtype,
)
elif w1_offset is not None:
# gmm1: gate_up_proj
hidden_states = torch_npu.npu_grouped_matmul(
x=[unquantized_hidden_states],
weight=[w1],
antiquant_scale=[w1_scale],
antiquant_offset=[w1_offset],
split_item=2,
group_list_type=group_list_type,
group_type=0,
group_list=group_list,
output_dtype=_output_dtype,
)[0]
dispose_tensor(unquantized_hidden_states)
# act_fn: swiglu
hidden_states = torch_npu.npu_swiglu(hidden_states)
# gmm2: down_proj
hidden_states = torch_npu.npu_grouped_matmul(
x=[hidden_states],
weight=[w2],
antiquant_scale=[w2_scale],
antiquant_offset=[w2_offset],
split_item=2,
group_list_type=group_list_type,
group_type=0,
group_list=group_list,
output_dtype=_output_dtype,
)[0]
else:
if w1_scale_bias is not None:
if group_list_type == 0:
group_list = torch.cat([group_list[:1], torch.diff(group_list, dim=0)])
group_list_type = 1
bias1 = [w1_scale_bias] if not fusion else w1_scale_bias
bias2 = [w2_scale_bias]
# TODO w4a8 scene: dynamic acquisition of dtype in the future
_output_dtype = torch.bfloat16
if _custom_gmm_swiglu_enabled(fusion, dynamic_eplb) and not use_mxfp_quant:
# gmm1: gate_up_proj & act_fn: swiglu
hidden_states, swiglu_out_scale, _ = torch.ops._C_ascend.grouped_matmul_swiglu_quant_weight_nz_tensor_list(
x=hidden_states,
weight=w1,
weight_scale=w1_scale,
x_scale=pertoken_scale,
group_list=cumsum_group_list(group_list, group_list_type, 0),
bias=bias1,
)
elif use_gmm_swiglu_quant_fusion:
hidden_states, swiglu_out_scale, _ = DeviceOperator.npu_grouped_matmul_swiglu_quant(
x=hidden_states,
weight=_require_single_tensor_for_swiglu_quant(w1, name="w1"),
group_list=cumsum_group_list(group_list, group_list_type, 0),
weight_scale=_require_single_tensor_for_swiglu_quant(w1_scale, name="w1_scale"),
x_scale=pertoken_scale,
bias=bias1,
use_mxfp_quant=use_mxfp_quant,
)
if quantized_hidden_states is not None:
dispose_tensor(quantized_hidden_states)
else:
w1_scale[0] = w1_scale[0].to(w2_scale[0].dtype)
# gmm1: gate_up_proj
hidden_states = torch_npu.npu_grouped_matmul(
x=[hidden_states],
weight=w1,
scale=w1_scale,
bias=bias1,
per_token_scale=[pertoken_scale],
split_item=2,
group_list_type=group_list_type,
group_type=0,
group_list=group_list,
output_dtype=_output_dtype,
)[0]
if quantized_hidden_states is not None:
dispose_tensor(quantized_hidden_states)
# act_fn: swiglu
if HAS_TRITON:
from vllm_ascend.ops.triton.activation.swiglu_quant import swiglu_quant
hidden_states, swiglu_out_scale = swiglu_quant(
hidden_states, group_list=group_list, group_list_type=group_list_type
)
else:
hidden_states = torch_npu.npu_swiglu(hidden_states)
hidden_states, swiglu_out_scale = torch_npu.npu_dynamic_quant(hidden_states)
# gmm2: down_proj
hidden_states = DeviceOperator.npu_grouped_matmul_gmm2(
hidden_states=hidden_states,
weight=w2,
weight_scale=w2_scale,
per_token_scale=swiglu_out_scale,
group_list=group_list,
group_list_type=group_list_type,
input_dtype=input_hidden_dtype,
act_quant_type=act_quant_type,
weight_quant_type=weight_quant_type,
scale_type=scale_type,
per_token_scale_type=per_token_scale_type,
use_bf16=use_bf16,
use_mxfp_quant=use_mxfp_quant,
bias=bias2,
fallback_output_dtype=_output_dtype,
)
return hidden_states
def unquant_apply_mlp(
hidden_states: torch.Tensor,
w1: torch.Tensor,
w2: torch.Tensor,
group_list: torch.Tensor,
w1_bias: torch.Tensor = None,
w2_bias: torch.Tensor = None,
activation: str | None = None,
group_list_type: int = 1,
topk_scales: torch.Tensor | None = None,
need_trans: bool = True,
) -> torch.Tensor:
if need_trans:
w1 = w1.transpose(1, 2)
w2 = w2.transpose(1, 2)
# In the small batch scenario, use _C_ascend.moe_grouped_matmul
if group_list.dim() == 2 and get_forward_context().num_tokens <= DeviceOperator.small_batch_gmm_batch_num:
gate_up_out = torch.ops._C_ascend.moe_grouped_matmul(
x=hidden_states,
weight=w1,
split_item=2,
group_list_type=group_list_type,
group_type=0,
group_list=group_list,
)[0]
else:
gate_up_out = torch_npu.npu_grouped_matmul(
x=[hidden_states],
weight=[w1],
bias=[w1_bias.to(dtype=torch.float32)] if w1_bias is not None else None,
split_item=2,
group_list_type=group_list_type,
group_type=0,
group_list=group_list,
)[0]
if activation == "swigluoai":
num_experts, _, hidden_size = w1.shape
gate_up_out = AscendSwigluOAIAndMul.swiglu_oai_forward(gate_up_out.view(-1, hidden_size))
else:
gate_up_out = torch_npu.npu_swiglu(gate_up_out)
if topk_scales is not None:
gate_up_out *= topk_scales
# In the small batch scenario, use _C_ascend.moe_grouped_matmul
if group_list.dim() == 2 and get_forward_context().num_tokens <= DeviceOperator.small_batch_gmm_batch_num:
hidden_states = torch.ops._C_ascend.moe_grouped_matmul(
x=gate_up_out,
weight=w2,
split_item=2,
group_list_type=group_list_type,
group_type=0,
group_list=group_list,
)[0]
else:
hidden_states = torch_npu.npu_grouped_matmul(
x=[gate_up_out],
weight=[w2],
bias=[w2_bias.to(dtype=torch.float32)] if w2_bias is not None else None,
split_item=2,
group_list_type=group_list_type,
group_type=0,
group_list=group_list,
)[0]
return hidden_states
def unified_apply_mlp(
hidden_states: torch.Tensor,
w1: torch.Tensor | list[torch.Tensor],
w2: torch.Tensor | list[torch.Tensor],
group_list: torch.Tensor,
w1_scale: list[torch.Tensor] | None = None,
w2_scale: list[torch.Tensor] | None = None,
activation: str | None = None,
w1_bias: torch.Tensor = None,
w2_bias: torch.Tensor = None,
dynamic_scale: torch.Tensor = None,
group_list_type: int = 1,
w1_scale_bias: torch.Tensor = None,
w2_scale_bias: torch.Tensor = None,
w1_offset: torch.Tensor | None = None,
w2_offset: torch.Tensor | None = None,
topk_scales: torch.Tensor | None = None,
with_quant: bool = False,
fusion: bool = False,
need_trans: bool = True,
dynamic_eplb: bool = False,
**kwargs,
) -> torch.Tensor:
"""
Unified MoE MLP entry.
Quant path is dispatched by DeviceOperator with explicit quant-type flags.
"""
if not with_quant:
return unquant_apply_mlp(
hidden_states=hidden_states,
w1=w1,
w2=w2,
w1_bias=w1_bias,
w2_bias=w2_bias,
activation=activation,
group_list=group_list,
group_list_type=group_list_type,
topk_scales=topk_scales,
need_trans=need_trans,
)
assert w1_scale is not None and w2_scale is not None
# TODO(linfeng): Current massive parameter passing is quite severe; parameter differences introduced by different
# quantization modes will be consolidated into a dataclass in a follow-up.
act_quant_type = kwargs.get("act_quant_type", torch.float8_e4m3fn)
weight_quant_type = kwargs.get("weight_quant_type", torch.float8_e4m3fn)
scale_type = kwargs.get("scale_type")
per_token_scale_type = kwargs.get("per_token_scale_type")
use_mxfp_quant = kwargs.get("use_mxfp_quant", False)
return quant_apply_mlp(
hidden_states=hidden_states,
w1=w1,
w1_scale=w1_scale,
w2=w2,
w2_scale=w2_scale,
group_list=group_list,
dynamic_scale=dynamic_scale,
group_list_type=group_list_type,
w1_scale_bias=w1_scale_bias,
w2_scale_bias=w2_scale_bias,
w1_offset=w1_offset,
w2_offset=w2_offset,
fusion=fusion,
dynamic_eplb=dynamic_eplb,
act_quant_type=act_quant_type,
weight_quant_type=weight_quant_type,
scale_type=scale_type,
per_token_scale_type=per_token_scale_type,
use_mxfp_quant=use_mxfp_quant,
use_bf16=kwargs.get("use_bf16", True),
)
Reference in New Issue View Git Blame Copy Permalink