### What this PR does / why we need it?
Refactors the Mixture-of-Experts (MoE) communication method selection
logic. The choice between all-gather, all-to-all, and mc2 is now
determined by expert parallel configuration, SoC version (A2/A3), and
token count for better performance.
### Does this PR introduce _any_ user-facing change?
None.
### How was this patch tested?
Added.
- vLLM version: v0.10.1.1
- vLLM main:
https://github.com/vllm-project/vllm/commit/eafa8dcde63d625350ed618db4dd1cbcbaae77a1
---------
Signed-off-by: Yizhou Liu <liu_yizhou@outlook.com>
### What this PR does / why we need it?
When both speculative decoding and aclgraph are applied, and
cudagraph_capture_sizes uses the default value, it will report that the
stream resources are insufficient.
### Does this PR introduce _any_ user-facing change?
### How was this patch tested?
- vLLM version: v0.10.1.1
- vLLM main:
https://github.com/vllm-project/vllm/commit/9c99e4871f12aa57072b3dfd14157018d30e327b
Signed-off-by: withHades <244036962@qq.com>
### What this PR does / why we need it?
AscendQuantizer/LLMQuantizer class is used to select quant method based
on quant config and some other arguments,
but it is more simple and clean replacing these classes with map. So i
remove them.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
ut and e2e test
- vLLM version: v0.10.1.1
- vLLM main:
https://github.com/vllm-project/vllm/commit/6997a25ac65ed6cc3c2be6d09ca45f633a345f63
Signed-off-by: 22dimensions <waitingwind@foxmail.com>
### What this PR does / why we need it?
In vllm version 0.10.1, a new KVOutputAggregator was added to the
executor, moving aggregation to the
executor(https://github.com/vllm-project/vllm/pull/19555). This caused
mooncake_connector to break. This change aims to fix this bug and also
adds a policy to forcibly release the KV cache when the prefill node
times out.
This PR is currently linked to a PR in vllm
(https://github.com/vllm-project/vllm/pull/23917). The vllm PR aims to
modify the finish and send count confirmation in heterogeneous TP
situations.
The reason for deleting many UTs is that a lot of communication codes
have been deleted, so the UT as a whole will appear more concise.
- vLLM version: v0.10.1.1
- vLLM main:
https://github.com/vllm-project/vllm/commit/fa4311d85f3a5f6d445f20429ca6d38122908eeb
---------
Signed-off-by: baxingpiaochong <771405853@qq.com>
### What this PR does / why we need it?
This PR ports #2312#2506#2531 to main branch.
Original implementation of torchair caching forces users to make
everything prepared, fix all the configuration and enable
`use_cached_npu_graph`, and it might cause some problems confusing to
understand and tackle for users. It is better to compile the graph twice
instead of reusing the old kvcaches and cached torchair graph. And the
extra duration time is acceptable. Additionally, this pr fixes a
recompilation problem of torchair graph mode caused by
`running_in_graph` variable in `AscendMLATorchairImpl`.
### Does this PR introduce _any_ user-facing change?
If users want to enabling torchair.cache_compile with high compilation
speed, it is recommended to enable both `use_cached_kv_cache_bytes` and
`use_cached_graph` in `torchair_graph_config`. Without
`use_cached_kv_cache_bytes`, we'll compile torchair computation graph
twice to avoid runtime error caused by configuration mismtaches (the
second compilation will be much faster). Additionally, we've made a
change to how the TORCHAIR_CACHE_HOME enviroment variable is utilized to
enhance safety and prevent accidental file deletion by adding a suffix
directory.
### How was this patch tested?
CI and e2e vllm serving pass.
- vLLM version: v0.10.1.1
- vLLM main:
https://github.com/vllm-project/vllm/commit/70549c1245c3eeb3706e3c09a9e18d702fbf705f
---------
Signed-off-by: linfeng-yuan <1102311262@qq.com>
See #2687 we notice that test_platform and test_vocab_parallel_embedding
is unstable, let's skip them first.
Signed-off-by: wangxiyuan <wangxiyuan1007@gmail.com>
### What this PR does / why we need it?
Fix long context seq accuracy problem for `GLM4.5`.
When `max_tokens=1000`, there is cyclic output problem like:
```bash
00 00 00 00 00 00 00 00 00 00 00 00 00 00
```
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
```python
import os
os.environ["VLLM_USE_MODELSCOPE"] = "True"
os.environ["VLLM_WORKER_MULTIPROC_METHOD"] = "spawn"
from vllm import LLM, SamplingParams
def main():
prompts = [
"Hello, my name is",
"The president of the United States is",
"The capital of France is",
"The future of AI is",
]
# Create a sampling params object.
sampling_params = SamplingParams(max_tokens=1000, temperature=0.0)
# Create an LLM.
llm = LLM(model="/root/.cache/modelscope/hub/models/ZhipuAI/GLM-4___5",
tensor_parallel_size=8,
enforce_eager=True,
trust_remote_code=True,
max_model_len=1024)
# Generate texts from the prompts.
outputs = llm.generate(prompts, sampling_params)
for output in outputs:
prompt = output.prompt
generated_text = output.outputs[0].text
print(f"Prompt: {prompt!r}, Generated text: {generated_text!r}")
if __name__ == "__main__":
main()
```
- vLLM version: v0.10.1.1
- vLLM main:
https://github.com/vllm-project/vllm/commit/0235103cbbdb511e6708aae600f759060a797c16
---------
Signed-off-by: Shanshan Shen <87969357+shen-shanshan@users.noreply.github.com>
Signed-off-by: shen-shanshan <467638484@qq.com>
### What this PR does / why we need it?
This PR enables `npu_moe_gating_top_k_softmax` when running quantized
MoE (such as W8A8). This op in fact makes no distinction between
quantized and non-quantized scenarios. Introducing this op reduces 3~4ms
for TPOT.
### Does this PR introduce _any_ user-facing change?
No.
### How was this patch tested?
- vLLM version: v0.10.1.1
- vLLM main:
https://github.com/vllm-project/vllm/commit/ce30dca5c44353f278dc114bd6f03b11700088eb
Signed-off-by: Angazenn <supperccell@163.com>
Refactor E2E CI to make it clear and faster
1. remove some uesless e2e test
2. remove some uesless function
3. Make sure all test runs with VLLMRunner to avoid oom error
4. Make sure all ops test end with torch.empty_cache to avoid oom error
5. run the test one by one to avoid resource limit error
- vLLM version: v0.10.1.1
- vLLM main:
https://github.com/vllm-project/vllm/commit/a344a5aa0a58cc1758d9721e848ce1f5ca4b6c7f
Signed-off-by: wangxiyuan <wangxiyuan1007@gmail.com>
### What this PR does / why we need it?
There are a lot of redundant codes related to moe here, and the
structure is not very clear.
We did the following things:
we have placed the relatively independent code related to apply_mlp into
a separate file;
removed the environment variables of alltoall_buffer and alltoall_seq.
Remove the code related to alltoall_buffer and alltoall_seq, and retain
the sole TokenDispatcher inheritance class.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
e2e&ut
- vLLM version: v0.10.1.1
- vLLM main:
https://github.com/vllm-project/vllm/commit/4071c76cf3cff46e14630f1e66cbf006b6eb51d3
---------
Signed-off-by: Pr0Wh1teGivee <calvin_zhu0210@outlook.com>
Signed-off-by: weijinqian_v1 <weijinqian@huawei.com>
Co-authored-by: weijinqian0 <12153182+weijinqian0@users.noreply.github.com>
### What this PR does / why we need it?
* **Unify execution paths:** Consolidates the quantized and
non-quantized execution paths into a single `fused_experts` function,
removing duplicated logic and making the control flow clearer and easier
to maintain.
* **W8A8 dynamic quantization:** Adds support for W8A8 dynamic
quantization inside the unified MoE kernel. Communication routines are
updated to correctly handle dynamic quantization scales for activations.
* **Weight pre-processing:** Prae-transpose the `w13` and `w2` weight
matrices (as implemented in PR #2025) so that quantized and
non-quantized models follow the same code path for the MoE gating,
up-projection, and down-projection operations.
* **All-to-all communication:** Adds an `all-to-all` collective
communication pattern. For large token counts on modern hardware,
`all-to-all` is more efficient than the previous `all-gather` strategy.
However, `all-to-all` is not really captured and replayed due to
multiple D2H operations which will trigger synchronization, and thus
raise error when capture graphs. We only use `all-to-all` when fallback
to `compiled_graph_for_general_shape`.
* **Dynamic communication selection:** The model runner now selects the
optimal MoE communication method (`mc2`, `allgather`, or `alltoall`) at
runtime based on token count and the Ascend SoC version.
* **Limitation:** `all-gather` is not yet supported for quantized
models, which means there is still something left to do on A2.
### Does this PR introduce _any_ user-facing change?
None.
### How was this patch tested?
No further test cases needed.
- vLLM version: v0.10.1.1
- vLLM main:
https://github.com/vllm-project/vllm/commit/d660c98c1b59580af97d6c7dd162c7c8894d40ed
---------
Signed-off-by: Yizhou Liu <liu_yizhou@outlook.com>
### What this PR does / why we need it?
1. quickfix mc2 operator error in aclgraph + ep<16 scenario to recover
CI, will be refactorred in the future
2. disable aclgraph when testing w8a8
### How was this patch tested?
CI passed with existing test.
- vLLM version: v0.10.1.1
- vLLM main:
https://github.com/vllm-project/vllm/commit/95089607fa307c5facfb9706ea919292fb56e78c
Signed-off-by: MengqingCao <cmq0113@163.com>