### What this PR does / why we need it?
Refactor MLP weight prefetch to consistency with MoE Model's prefetching
in terms of code and usage.
Environments VLLM_ASCEND_ENABLE_PREFETCH_MLP,
VLLM_ASCEND_MLP_DOWN_PREFETCH_SIZE and
VLLM_ASCEND_MLP_GATE_UP_PREFETCH_SIZE is removed, usage as following:
--additional-config '{"weight_prefetch_config": { "enabled": true,
"prefetch_ratio": {"mlp": { "gate_up": 1.0, "down": 1.0} }}}'
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
- vLLM version: v0.14.1
- vLLM main:
dc917cceb8
---------
Signed-off-by: leo-pony <nengjunma@outlook.com>
### What this PR does / why we need it?
1. Implement a **high-performance Triton custom kernel** for the rotary
position embedding (RoPE) operator on **Ascend NPU** platform
2. Fix critical bugs in the Triton RoPE kernel registration and
invocation process: including incorrect fake impl function name
matching, wrong torch ops namespace for kernel call, missing self
parameter in cos/sin slice fetching, and syntax errors in function type
annotations.
3. Achieve **extreme performance optimization** for the core RoPE
operator: the single inference latency is reduced from **57.1 μs** to
**9 μs**, with **6.34x performance improvement** and **84.24% latency
reduction**.
4. The RoPE operator is a **hot path** that is executed in every
transformer layer during LLM inference, the optimization will directly
reduce the overall inference latency and improve the throughput of LLM
serving on Ascend NPU.
5. Keep full backward compatibility: the Triton kernel is enabled only
when `HAS_TRITON=True`, and automatically fall back to the original
Ascend NPU native implementation if Triton is not available, no
functional regression.
### Does this PR introduce _any_ user-facing change?
**NO**
- No changes to any public APIs, interfaces or inference behaviors of
vLLM.
- No impact on the text generation quality and correctness of the large
model.
- The optimization is transparent to end users, only the inference speed
(latency/throughput) is improved without any functional change.
### How was this patch tested?
1. **Environment Validation**: Tested on Ascend NPU platform with
vLLM-Ascend framework, Triton library installed and enabled
(`HAS_TRITON=True`).
2. **Kernel Registration Test**: Verified the Triton RoPE kernel
(`rope_forward_triton`) is successfully registered to
`torch.ops._C_ascend` namespace without any
`ValueError/NameError/SyntaxError`.
3. **Functional Correctness Test**: Run large model (GLM4/MoE) inference
on the Ascend NPU platform, the generated text content is **completely
correct** (no garbled text, no logical errors), consistent with the
original implementation.
4. **Performance Benchmark Test**: Measure the single execution latency
of the RoPE operator before/after optimization, confirm the latency is
stably reduced from 57.1 μs to 9 μs, the performance gain is valid and
stable.
5. **Fallback Mechanism Test**: Manually disable Triton
(`HAS_TRITON=False`), verify the code correctly falls back to the
original Ascend NPU native RoPE implementation, no service crash and
normal inference.
6. **Compatibility Test**: Test with different tensor shapes/sizes of
query/key, all cases work correctly with the Triton kernel, no shape
mismatch error.
- operator supply by Hexiang Wang
- vLLM version: v0.13.0
- vLLM main:
11b6af5280
---------
Signed-off-by: ZCG12345 <2097562023@qq.com>
### What this PR does / why we need it?
#5051 only implement a basic framework for model runner v2, but there
are still some bugs for e2e functionality, this PR aim to enable basic
functionality.
model runner v2 plans:
https://github.com/vllm-project/vllm-ascend/issues/5208
- vLLM version: release/v0.13.0
- vLLM main:
ad32e3e19c
---------
Signed-off-by: Ronald1995 <ronaldautomobile@163.com>
### What this PR does / why we need it?
Remove unnecessary attributes from set_ascend_forward_context
1.prefetch_stream
2.weight_prefetch_method
### Does this PR introduce _any_ user-facing change?
no
### How was this patch tested?
- vLLM version: v0.12.0
- vLLM main:
ad32e3e19c
Signed-off-by: Wang Kunpeng <1289706727@qq.com>
### What this PR does / why we need it?
- Renames the MoE comm enum value `MoECommType.FUSED_ALLTOALL` to
`MoECommType.FUSED_MC2` and updates all call sites.
- Updates `select_moe_comm_method` to optionally select `FUSED_MC2` on
Ascend A3 when:
- `enable_expert_parallel=True`
- quantization is `w8a8_dynamic`
- `EP <= 16`
- `dynamic_eplb` is disabled
- `is_mtp_model = False`
- Replaces the old “fused all-to-all” comm implementation with
`FusedMC2CommImpl`, using `TokenDispatcherWithMC2` /
`PrepareAndFinalizeWithMC2` and `dispatch_ffn_combine`.
- vLLM version: v0.12.0
- vLLM main:
ad32e3e19c
---------
Signed-off-by: Chen Chen <0109chenchen@gmail.com>
### What this PR does / why we need it?
1. In addition to
[#4168](https://github.com/vllm-project/vllm-ascend/pull/4168),
[#5011](https://github.com/vllm-project/vllm-ascend/pull/5011)， this PR
adds two more pattern for AddRmsnormQuant with SP enabled. The key
difference is to insert an additional `maybe_all_gather_and_maybe_unpad`
between `addrmsnorm` and `quantize`.
2. This PR also introduce another api `torch.ops.vllm.quantize`, so that
we pass `input_scale` and `input_scale_reciprocal` at the same time.
This is because `npu_add_rms_norm_quant` and `npu_quantize` requires
different `div_mode`. To avoid introducing additional reciprocal
calculation in runtime, we have to pass both of them to quantize api.
3. Removes redundant `AscendQuantRmsnorm`.
- vLLM version: v0.12.0
- vLLM main:
ad32e3e19c
---------
Signed-off-by: Angazenn <supperccell@163.com>
### What this PR does / why we need it?
Related to #4084. Before we add the patches temporarily for making
`set_forward_context` patched by `set_ascend_forward_context` in the
function `_process_image_input` and `_process_video_input` of
`Qwen2.5-VL` and `Qwen2.5-Omni` models. After removing these patches, I
met the `AttributeError` for `ForwardContext` missing
`prefetch_mlp_enabled`. So we need to add the defensive check for
`prefetch_mlp_enabled`.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
```
vllm serve Qwen/Qwen2.5-VL-7B-Instruct \
--max-model-len 30000 \
--max-num-batched-tokens 50000 \
--max-num-seqs 30 \
--no-enable-prefix-caching \
--trust-remote-code \
--dtype bfloat16
```
```
{"id":"chatcmpl-b66d8acb76905c49","object":"chat.completion","created":1765796863,"model":"/root/.cache/modelscope/hub/models/Qwen/Qwen2.5-VL-7B-Instruct","choices":[{"index":0,"message":{"role":"assistant","content":"The text in the illustration reads \"TONGYI Qwen.\"","refusal":null,"annotations":null,"audio":null,"function_call":null,"tool_calls":[],"reasoning":null,"reasoning_content":null},"logprobs":null,"finish_reason":"stop","stop_reason":null,"token_ids":null}],"service_tier":null,"system_fingerprint":null,"usage":{"prompt_tokens":73,"total_tokens":88,"completion_tokens":15,"prompt_tokens_details":null},"prompt_logprobs":null,"prompt_token_ids":null,"kv_transfer_params":null}
```
- vLLM version: v0.12.0
- vLLM main:
ad32e3e19c
---------
Signed-off-by: gcanlin <canlinguosdu@gmail.com>
Co-authored-by: wangxiyuan <wangxiyuan1007@gmail.com>
### What this PR does / why we need it?
This PR introduces the Ascend implementation of the
`dispatch_ffn_combine` kernel and wires it into the vLLM-Ascend runtime,
together with follow‑up fixes to ensure the kernel builds and runs
correctly in CI.
- Add full host and device implementation of the `dispatch_ffn_combine`
kernel under `csrc/dispatch_ffn_combine`, including tiling logic, MOE
routing helpers, and kernel utilities for quantized FFN dispatch.
- Integrate the new kernel with the PyTorch binding
(csrc/torch_binding.cpp, csrc/torch_binding_meta.cpp) and the Ascend
runtime (vllm_ascend/ascend_forward_context.py,
vllm_ascend/worker/model_runner_v1.py).
- Extend fused MoE communication and token dispatch support in
`vllm_ascend/ops/fused_moe`, adding methods/utilities needed by the new
dispatch path.
- Update quantization logic in vllm_ascend/quantization/w8a8_dynamic.py
to support the new FFN dispatch flow.
- Fix kernel build issues by adjusting `csrc/build_aclnn.sh`, CMake
configuration, and include/namespace usage in the new kernel files.
- Add an end‑to‑end nightly test
`tests/e2e/nightly/ops/test_dispatch_ffn_combine.py` and helper
utilities in `vllm_ascend/utils.py` to validate the new kernel.
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
- vLLM version: v0.12.0
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.12.0
---------
Signed-off-by: mojave2 <chenchen145@huawei.com>
Co-authored-by: wangxiyuan <wangxiyuan1007@gmail.com>
### What this PR does / why we need it?
Support shared expert DP for deepseek_mtp feature.
`shared_expert_dp` requires `SP==True`, with corresponding parameter
restrictions.
Previously, due to the coupling between `shared_expert_dp` and torchair,
and the removal of `deepseek_mtp` in vllm_ascend, shared expert dp of
deepseek_mtp was temporarily removed.
Currently, by performing the `reduce_scatter` on the input of
deepssek_mtp in `mtp_proposer.py`, we ensure that it matches the
dimensions of `input_embedding`, and then perform the `all_gather` on
the output of mtp.
### How was this patch tested?
baseline:
<img width="1184" height="692" alt="image"
src="https://github.com/user-attachments/assets/9680d53a-7b1d-481a-accc-b8f3dae2b9e3"
/>
enable shared_expert_dp and multistream_overlap_shared_expert:
<img width="1167" height="687" alt="image"
src="https://github.com/user-attachments/assets/2531d06b-dfda-4e24-8628-6f4b0f677ddc"
/>
TPOT: 48ms -> 45.4ms
Average TPS per rank: 117.6 -> 126.1
- vLLM version: v0.11.2
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.2
---------
Signed-off-by: chenmenglong <chenmenglong1@huawei.com>
Signed-off-by: zengran <zengran2@huawei.com>
Co-authored-by: zengran <zengran2@huawei.com>
There is a lot hack code for v0.11.0, which makes the code hard to
upgrade to newer vLLM version. Since v0.11.0 will release soon. Let's
drop v0.11.0 support first. Then we'll upgrade to v0.11.2 soon.
- vLLM version: v0.11.0
- vLLM main:
2918c1b49c
Signed-off-by: wangxiyuan <wangxiyuan1007@gmail.com>
### What this PR does / why we need it?
move quant before allgather in Allgather EP, rely on
https://github.com/vllm-project/vllm-ascend/pull/3334
Deepseek R1 W8A8 performance on A2 with
`HCCL_ALGO="level0:NA;level1:pipeline"`:
| Seq length | Mean TTFT (ms) main | Mean TTFT (ms) this PR |
|----------|----------|----------|
| 4k | 375.21 | 364.99 |
| 16k | 1465.23 | 1421.75 |
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
- vLLM version: v0.11.0
- vLLM main:
83f478bb19
---------
Signed-off-by: realliujiaxu <realliujiaxu@163.com>
### What this PR does / why we need it?
This PR refactors SequenceRowParallelOp forward. In order to further
expand the operator inclusion scope in dynamic judgment scenarios, this
PR customizes the entire matmul computation and communication as a
custom operator masking. With this refactor, it will support directly
writing code such as common operation fusion into the
`SequenceRowParallelOp` class's member function `matmul_and_reduce`,
without the need to register more redundant custom masking operators.
### How was this patch tested?
CI passed with existing test.
Signed-off-by: rjg-lyh <1318825571@qq.com>
### What this PR does / why we need it?
This reverts commit
bf87606932.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
E2E vllm serving with `enable_shared_expert_dp: true` in eager mode as
before.
- vLLM version: v0.11.0rc3
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.0
Signed-off-by: linfeng-yuan <1102311262@qq.com>
### What this PR does / why we need it?
shared expert dp for deepseek and deepseek_mtp, could be combined with
sp to improve performance.
### How was this patch tested?
- vLLM version: v0.11.0rc3
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.0
---------
Signed-off-by: zhaozx-cn <zhaozx2116@163.com>
Co-authored-by: realliujiaxu <realliujiaxu@163.com>
This PR will accomplish the following tasks:
**optimize SP**
In the old version implementation, the first layer was all_reduce, which
used rms to split chunks. We changed it to perform reduce_scatter on the
embedding side, replace one all_reduce operation and one chunk with one
reduce_scatter operation.
**Support qwen3 next**
Since Qwen3 Next includes a linear attention module, the prefix name of
this module cannot take effect directly.
- vLLM version: v0.11.0rc3
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.0
---------
Signed-off-by: weijinqian_v1 <weijinqian@huawei.com>
Co-authored-by: weijinqian_v1 <weijinqian@huawei.com>
### What this PR does / why we need it?
- Refacotr and integrate a unified `WeightPrefetchMethod`
- Integrate `qkv_proj.weight` and `o_proj.weight` in quantized Attention
modules
- Prefetching these weights ahead of matmul-like operators imporves
performance by reducing L2 cache transfer latency
### Does this PR introduce _any_ user-facing change?
Add a new config in `--additional-config` for configuration:
```json
{
"weight_prefetch_config": {
"enabled": false,
"prefetch_ratio": {
"attn": {
"qkv": 1.0,
"o": 1.0,
},
},
},
}
```
This feature is enabled by default, and can be disabled through this
configuration
### How was this patch tested?
- vLLM version: v0.11.0
---------
Signed-off-by: yuzhup <15705211260@163.com>
Signed-off-by: zhoux77899 <zhouxiang100@huawei.com>
Co-authored-by: yuzhup <15705211260@163.com>
This PR fixes accuracy problem of aclgraph on A2. The problem is
introduced by PR #2980, which makes the `all_reduce` of shared_experts
exposed to torch dynamo. This PR moves all the codes into forward_impl
to shiled from torch dynamo.
- vLLM version: v0.10.2
- vLLM main:
17b4c6685c
---------
Signed-off-by: whx-sjtu <2952154980@qq.com>
What this PR does / why we need it?
there are two sets of sp implementations for moe and dense models. One
is called sequence_parallelism, and the other is flashcomm_v1.
We did the following things：
Merge two sets of code with the same implementation into one.
Remove the implementation of sequence_parallelism, as this solution
cannot support aclgraph.
Does this PR introduce any user-facing change?
No
How was this patch tested?
e2e&ut
- vLLM version: v0.10.2
- vLLM main:
f225ea7dd9
---------
Signed-off-by: weijinqian_v1 <weijinqian@huawei.com>
Co-authored-by: weijinqian_v1 <weijinqian@huawei.com>
### What this PR does / why we need it?
This PR removed the redundant log prints in register_custom_ops.py, in
order to make output clear.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
CI passed with new added/existing test.
- vLLM version: v0.10.2
- vLLM main:
9607d5eb44
Signed-off-by: rjg-lyh <1318825571@qq.com>
### What this PR does / why we need it?
This PR fixed the bug in register_custom_ops without forward_context. We
set try-except to consider this situation.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
CI passed with new added/existing test.
- vLLM version: main
- vLLM main:
7920de0a2a
Signed-off-by: rjg-lyh <1318825571@qq.com>
### What this PR does / why we need it?
This PR prefetchs the weight of mlp layers in Qwen Dense Models to
optimize the performance in Decode phase mainly.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
CI passed with new added/existing test.
- vLLM version: main
- vLLM main:
a1213fae5f
Signed-off-by: rjg-lyh <1318825571@qq.com>
Co-authored-by: Shuming19 <313093131@qq.com>
### What this PR does / why we need it?
Flashcomm_v1 optim in Qwen Dense Models.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
CI passed with new added/existing test.
- vLLM version: v0.10.1.1
- vLLM main:
5e537f45b4
Co-authored-by: 1024daniel <xxltju324@gmail.com>
