### What this PR does / why we need it?
https://github.com/vllm-project/vllm-ascend/pull/2849 moves the
implementation of `shared_expert_dp` to torchair deepseek_modeling.
However, the calling of `set_forward_context` with `enforce_eager` and
`shared_expert_dp` falls back to the implementation of
model_runner_v1.py and set the global attn_metadata as a dictionary. It
leads to a RuntimerError when attn_metadata is got from the forward
context and used in torchair_deepseek_v2.py. This PR fixes this problem
by introducing the transformation of attn_metadata in this file.
Note that current E2E testing lacks the case of deepseek with
`shared_expert_dp`. We need to add an ST with `shared_expert_dp` in
testing workflow.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
e2e vllm serving with `enable_shared_expert_dp: true` passed.
- vLLM version: v0.10.2
- vLLM main:
https://github.com/vllm-project/vllm/commit/de3e53a75ba9f31f446926911b7c44561af3b2ee
Signed-off-by: linfeng-yuan <1102311262@qq.com>
### Motivation
Currently dynamically experts balancing would stop-the-world.
Asynchronously expert load balancing would be better without flowing
problems:
Host-bound latency:
There are many cpu operations during EPLB such as
eplb-algorithm、creating p2p ops、and log2phy expert converting would
spend long cpu time, as ~1s.
Communication latency: The transfer time would cost much in the
situation without nvlink. As the weight of an expert maybe transfer to
multiple new positions, thus N times send/recv for one expert, with
result long latency. We had tested that batch_isend_irecv cost more
100ms for 16 experts weight transmission in A2 server of ascend.
SwiftBalancer would not stop-the-world anymore, in out test on NPU 1~2ms
cost for each layer while benefit 5ms-8ms decode latency with ep_size =
64.
The following updates have been made:
1、expert distribution recording with lower cost.
2、async cpu computing for eplb algo and other python operator.
3、new eplb algo with less expert rebalancing while almost the same
effect.
### Proposed Change
We will gradually migrate the EPLB logic to the VLLM community and
implement a generalized design. Relevant RFC:
https://github.com/vllm-project/vllm/issues/22246
The overall workflow involves:
<img width="801" height="302"
alt="474430541-23b06f58-23bc-44a3-a1be-00f268aeb15c"
src="https://github.com/user-attachments/assets/1d73a459-1b23-4b0a-812a-bf0a75debfed"
/>
1. Record experts distribution during forward. We using expert_token_num
after disptach instead of topk_ids, thus we got much smaller tensor
shape to reduce cost of hbm recording and add-operator.
2. Do all-gather for experts distribution. Using all-gather instead of
all-reduce as less traffic volume.
3. Wake up eplb worker process with experts distribution when
num_iterations comes. Run eplb algorithm in eplb worker.
4. Generate p2p send/recv ops and other operator such as log2phy would
cost long cpu time.
5. Lanch ibatch_send_recv in async_stream before forward.
6. After forward, wait for the ibatch_send_recv finish, then do uapte
expert map and expert weights.
### Co-author
Co-authored-by: raindaywhu raindaywhu@raindaywhu@ 163.con
Co-authored-by: njuyuan yuanjl19@smail.nju.edu.cn
Co-authored-by: qmkakaxi wjh1594260677@qq.com
Co-authored-by: Skywalker-EP 173723846@qq.com
- vLLM version: v0.10.2
- vLLM main:
https://github.com/vllm-project/vllm/commit/567939953b7a9cb0ded6bf0bb21a76917b8fed97
---------
Signed-off-by: offline0806 <z00858301@china.huawei.com>
Co-authored-by: offline0806 <z00858301@china.huawei.com>
### What this PR does / why we need it?
1. In memory of #2509, Fix mtp torchair in pd Disaggregation scenario
2. fix mla bug in SpecDecoding Scenario, since num_decodes !=
num_decode_tokens
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
- vLLM version: v0.10.2
- vLLM main:
https://github.com/vllm-project/vllm/commit/5206ab20ba7477e5457c2e64469590d548fa15e6
Signed-off-by: xuyexiong <xuyexiong@huawei.com>
### What this PR does / why we need it?
1. Replace prepare/finalize operation in fused_moe.py by
moe_comm_method.prepare()/finalize()
2. Replace unified_fused_experts by moe_comm_method.fused_experts() in
fused_moe.py/w8a8_dynamic.py/w4a8_dynamic.py
3. Add calling _select_moe_comm_method in spec-decode proposers.
4. Currently, w4a8_dynamic does not support gatherep, use all2allv
instead.
5. Remove redundant code.
### Does this PR introduce _any_ user-facing change?
AllgatherEP switch is disabled in aclgraph/eager mode, just follow the
rules in modelrunner_v1._select_moe_comm_method()
### How was this patch tested?
e2e & ut
- vLLM version: v0.10.2
- vLLM main:
https://github.com/vllm-project/vllm/commit/7f6f2c1182b4f6e23d9d10715e25f60fbfd8e6ac
Signed-off-by: Pr0Wh1teGivee <calvin_zhu0210@outlook.com>
Co-authored-by: weijinqian0 <12153182+weijinqian0@users.noreply.github.com>
### What this PR does / why we need it?
Update max_tokens and prompt in qwen3 online doc
Before:
```
"'max_tokens' or 'max_completion_tokens' is too large: 4096. This model's maximum context length is 4096 tokens and your request has 18 input tokens (4096 > 4096 - 18). None"
```
After:
```
curl http://localhost:8000/v1/chat/completions -H "Content-Type: application/json" -d '{
"model": "/root/.cache/modelscope/hub/models/Qwen-SGlang/Qwen3-Next-80B-A3B-Instruct",
"messages": [
{"role": "user", "content": "Who are you?"}
],
"temperature": 0.6,
"top_p": 0.95,
"top_k": 20,
"max_tokens": 32
}'
.{"id":"chatcmpl-8ddbd65c9ddc405397219a6792feb9a0","object":"chat.completion","created":1757985049,"model":"/root/.cache/modelscope/hub/models/Qwen-SGlang/Qwen3-Next-80B-A3B-Instruct","choices":[{"index":0,"message":{"role":"assistant","content":"Hello! I am Qwen, a large-scale language model independently developed by the Tongyi Lab under Alibaba Group. I am designed to assist you in generating various","refusal":null,"annotations":null,"audio":null,"function_call":null,"tool_calls":[],"reasoning_content":null},"logprobs":null,"finish_reason":"length","stop_reason":null,"token_ids":null}],"service_tier":null,"system_fingerprint":null,"usage":{"prompt_tokens":12,"total_tokens":44,"completion_tokens":32,"prompt_tokens_details":null},"prompt_logprobs":null,"prompt_token_ids":null,"kv_transfer_params":null}
```
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
- Manually test on my local env
- CI passed
Signed-off-by: Yikun Jiang <yikunkero@gmail.com>
### What this PR does / why we need it?
On main, AscendScheduler does not support Multimodels, becuse of lacking
of scheduled_encoder_inputs which is need on multimodels inference
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
vLLM version: main@93e28e6862669e3b5cf47cea9f782a65ec47e155
- vLLM version: v0.10.2rc2
- vLLM main:
https://github.com/vllm-project/vllm/commit/15b8fef453b373b84406207a947005a4d9d68acc
---------
Signed-off-by: fan2956 <zhoufan53@huawei.com>
Co-authored-by: zhoufan2956 <zhoufan2956@163.com>
### What this PR does / why we need it?
This PR is used to adapt the hostname format for Mooncake when using
adxl. When Mooncake uses adxl, it is necessary to set
```USE_ASCEND_DIRECT``` to True in the file
```/Mooncake/mooncake-common/common.cmake``` during compilation. The
mooncake_connector obtains this config by calling
```vllm_config.kv_transfer_config.get_from_extra_config```, determines
whether Mooncake is using adxl, and selects the corresponding hostname
format.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
By CI.
- vLLM version: main
- vLLM main:
https://github.com/vllm-project/vllm/commit/d21a36f5f9569949dd6c313deed609d77e393850
---------
Signed-off-by: nwpu-zxr <zhouxuerong2@huawei.com>
**Background:**
There are two principles about operator registration in PyTorch
- The same namespace can be only registered once by `TORCH_LIBRARY`
- The operator signatures can be only registered once by `def`
Considering that all custom operators defined in the current repo are
only used by Ascend, instead of defining a common operator schema by
vLLM, all accelerators then follow this operator schema and complete the
implementation based on their respective hardware, which is conducive to
functional abstraction.
Therefore, we can rename the operator registration namespace to an
Ascend-specific namespace(**_C_ascend**).
Related ISSUE: https://github.com/vllm-project/vllm-ascend/issues/2742
- vLLM version: main
- vLLM main:
https://github.com/vllm-project/vllm/commit/f592b3174b39a7aeac52432d66d76e89ff0a80b4
Signed-off-by: FFFrog <ljw1101.vip@gmail.com>
### What this PR does / why we need it?
This PR enforces the forcible disabling of the chunked prefill feature
in Non-MLA models, as the performance of operators supporting this
functionality is currently suboptimal. Unless the user has enabled
chunked prefill in the ascend_scheduler_config, we would allow this
feature.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
CI passed with new added/existing test.
Related: https://github.com/vllm-project/vllm-ascend/pull/2659
- vLLM version: main
- vLLM main:
https://github.com/vllm-project/vllm/commit/d21a36f5f9569949dd6c313deed609d77e393850
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:
https://github.com/vllm-project/vllm/commit/7920de0a2af7c663efae29856f1bac061a553b58
Signed-off-by: rjg-lyh <1318825571@qq.com>
### What this PR does / why we need it?
1. Move ops/comm_utils to ops/moe/comm_utils
2. Move distributed/tensor_parallel/gather_from_sequence_parallel_region
to ops/moe/comm_utils
3. Delete distributed/tensor_parallel
### Does this PR introduce _any_ user-facing change?
no
### How was this patch tested?
e2e & ut
- vLLM version: main
- vLLM main:
https://github.com/vllm-project/vllm/commit/a1213fae5f8fe624c4b954768d23a86e384abdb1
---------
Signed-off-by: wuweiqiang24 <1005334931@qq.com>
Signed-off-by: wuweiqiang24 <wuweiqiang11@huawei.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:
https://github.com/vllm-project/vllm/commit/a1213fae5f8fe624c4b954768d23a86e384abdb1
Signed-off-by: rjg-lyh <1318825571@qq.com>
Co-authored-by: Shuming19 <313093131@qq.com>
This PR is based on top of
[#23569](https://github.com/vllm-project/vllm/pull/23569) and
[#24219](https://github.com/vllm-project/vllm/pull/24219).
### What this PR does / why we need it?
This PR allows the model runner to function asynchronously when using
async scheduling. This allows full overlap of the cpu operations
(including prepare_inputs) and the model forward pass. This diff is
functional and does not support speculative decoding, PP, or guided
decoding.
Expected speedup is 5-10% over the current async scheduling.
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
server
```
python -m vllm.entrypoints.openai.api_server --model=Qwen3-32B\
--trust-remote-code --enforce-eager \
--distributed-executor-backend=mp \
-tp=4 \
--port 8006 \
--max-model-len 32000 \
--block-size 128 \
--gpu-memory-utilization 0.99
```
client
```
python $TEST_PY --backend vllm --trust-remote-code --model Qwen3-32B \
--dataset-name random --random-input-len 2048 --random-output-len 2048 \
--ignore-eos\
--num-prompts 48 --max-concurrency 48 --request-rate inf --temperature 0 \
--metric-percentiles 90 --base-url http://localhost:8006 --save-result \
--result-dir $PROFILER_DIR
```
benchmark test based on Qwen3-32B TPOT result:
||forward async| scheduler async |sync|
|-|-|-|-|
|avg|41.73|41.86|44.20|
|improve0|0.3%|0|0|
|improve1|5.58%|0|0|
benchmark test based on Qwen2___5-VL-7B-Instruct TPOT result:
||forward async|sync|
|-|-|-|
|avg|23.22|29.16|
|improve|20.3%|0|
- vLLM version: main
- vLLM main:
https://github.com/vllm-project/vllm/commit/e93f4cc9e37484009f74e15d3111a1f335c532a5
Signed-off-by: jiangpeng36 <jiangpeng36@huawei.com>
Signed-off-by: Ronald1995 <ronaldautomobile@163.com>
Co-authored-by: jiangpeng36 <jiangpeng36@huawei.com>
Co-authored-by: Ronald1995 <ronaldautomobile@163.com>
### What this PR does / why we need it?
Currently, when executing to the Linear layer of the model in
vLLM-Ascend, the weights input format is ND in unquantized case and
skipped ascend case, which is slower than FRACTAL_NZ.
This PR supplements the execution logic for Linear layer. When
VLLM_ASCEND_ENABLE_MLP_OPTIMIZE=1 and CANN version is 8.3, the weights
of the Linear layer will be converted to FRACTAL_NZ, in both unquantized
case and skipped ascend case.
- vLLM version: main
- vLLM main:
https://github.com/vllm-project/vllm/commit/267c80d31f6b77092a5d5903da64556ac15c4d4d
Signed-off-by: anon189Ty <Stari_Falcon@outlook.com>
### What this PR does / why we need it?
LLMdatadist connector adapt the distributed KV aggregation for the main
branch. Change the P node from returning "finish sending" only when TP0
responds to returning "finish sending" as soon as each NPU receives it.
The D node will send a finish receive signal to the corresponding tp
rank of the P node.
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
gsm8k test
2*A3 1P 1D
P: dp2 tp8 D:dp 4 tp4
P: dp2 tp8 D:dp 2 tp8
- vLLM version: main
- vLLM main:
https://github.com/vllm-project/vllm/commit/cc99baf14dacc2497d0c5ed84e076ef2c37f6a4d
Signed-off-by: liziyu <liziyu16@huawei.com>
### What this PR does / why we need it?
1. Solved the problem that in the Qwen3 Moe model case, opening DP would
use an extra stream, causing ACLgraph sizes capture error
2. After experimentation, it was found that in many cases, some
operators would occupy more streams than expected. Therefore, the buffer
area for streams in ACLgraph was not large enough. After discussion,
extra 120 streams were added as buffer.
### Does this PR introduce _any_ user-facing change?
no
### How was this patch tested?
ut
- vLLM version: main
- vLLM main:
https://github.com/vllm-project/vllm/commit/0ae43dbf8cb28a299ae724fc742b0c5bcddea868
Signed-off-by: lilinsiman <lilinsiman@gmail.com>
**Background:**
A dynamic library named vllm_ascend_C.so will be generated when
compiling vLLM-Ascend, so when import vllm_ascend.vllm_ascend_C in
python, the interperter will search vllm_ascend_C.so and try to find a
external symbol named PyInit_vllm_ascend_C which is provided in
csrc/camem_allocator.cpp.
**Conclusion:**
The PyInit__C is redundent.
- vLLM version: v0.10.1.1
- vLLM main:
https://github.com/vllm-project/vllm/commit/717fc00e98c18d183f9141393b4185ddfb6606b3
Signed-off-by: FFFrog <ljw1101.vip@gmail.com>
### What this PR does / why we need it?
Register AscendMultiHeadLatentAttention as CustomOP, following vllm changes
### Does this PR introduce _any_ user-facing change?
N/A
### How was this patch tested?
CI passed with new added/existing test.
- vLLM version: main
- vLLM main:
https://github.com/vllm-project/vllm/commit/b23fb78623f0019854af7a820d398cba9a316677
---------
Signed-off-by: Icey <1790571317@qq.com>