### What this PR does / why we need it?
This PR add `qkv_rmsnorm_rope` operator and introduces a graph fusion
pass for `qknorm_rope` operations. The implementation includes a new
configuration flag, a pattern matching pass using
`torch._inductor.pattern_matcher`, and a custom Triton kernel for the
fused operation.
Co-authored-by: Angazenn
[supperccell@163.com](mailto:supperccell@163.com)
### Does this PR introduce _any_ user-facing change?
Yes, add new additional_config
- vLLM version: v0.12.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/ad32e3e19ccf0526cb6744a5fed09a138a5fb2f9
---------
Signed-off-by: wxsIcey <1790571317@qq.com>
### What this PR does / why we need it?
This PR adds triton-ascend installation to nightly e2e single card
environment.
Signed-off-by: whx-sjtu <2952154980@qq.com>
### What this PR does / why we need it?
The name of the smoke test file for DeepSeek EPLB has been changed, but
the name in the script hasn't been updated. Fix this bug.
Signed-off-by: shenchuxiaofugui <1311027364@qq.com>
What this PR does / why we need it?
a2 kernel aclnn interface extern "C" fix
Does this PR introduce any user-facing change?
No
How was this patch tested?
vLLM version: v0.12.0
Signed-off-by: tongrunze <t00574058@china.huawei.com>
Co-authored-by: tongrunze <t00574058@china.huawei.com>
### What this PR does / why we need it?
When use mtp, full decdoe only and async_scheduling together, finding a
input err for FIA ops due to the non-increasing input
of the 'actual_seq_lengths'. This bug is caused by the filling the
variable ‘query_start_loc’. We need to fill the query_start_loc' s end
by the 'cu_num_tokens' instead of '-1'
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
- vLLM version: v0.12.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/ad32e3e19ccf0526cb6744a5fed09a138a5fb2f9
Signed-off-by: hust17yixuan <303660421@qq.com>
### What this PR does / why we need it?
this pr removes apply_gramme in npu_model_runner. we change logits to
cpu, and do the same thing with gpu_model_runner.
it may change the performance, we will change it after torch.compile is
supported with npu inductor
- vLLM version: v0.12.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/ad32e3e19ccf0526cb6744a5fed09a138a5fb2f9
---------
Signed-off-by: zhenwenqi2024 <zhenwenqi_2022@qq.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:
https://github.com/vllm-project/vllm/commit/ad32e3e19ccf0526cb6744a5fed09a138a5fb2f9
---------
Signed-off-by: gcanlin <canlinguosdu@gmail.com>
Co-authored-by: wangxiyuan <wangxiyuan1007@gmail.com>
### What this PR does / why we need it?
The fused alltoall operator itself was not designed or implemented to
handle the scenario where tensors are lists, but the weights for dynamic
load balancing are in list form.
Therefore, we have disabled this operator when using dynamic load
balancing.
- vLLM version: v0.12.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/ad32e3e19ccf0526cb6744a5fed09a138a5fb2f9
Signed-off-by: shenchuxiaofugui <1311027364@qq.com>
### What this PR does / why we need it?
This PR introduces the initial integration of **UCM (Unified Cache
Management)** into the vllm-ascend distributed KV-cache system.
Specifically, it adds:
- A new `UCMConnector` implementation under the distributed KV-transfer
framework.
- Support for offloading KV-cache blocks to external UCM backends (DRAM
/ NFS / Localdisk), depending on UCM configuration).
- Integration with vLLM V1 KV connector interface, including metadata
handling and role registration.
**Why it is needed:**
- UCM provides a unified, high-performance storage layer for KV-cache
externalization.
- This enables vllm-ascend to support out-of-core KV-cache workloads,
improve memory efficiency, and leverage hardware-accelerated storage
paths (RDMA / NFS / hybrid modes).
- This connector is a required component to allow future work on
multi-node inference + UCM-based scaling.
---
### Does this PR introduce _any_ user-facing change?
Yes, but limited:
- A new `kv_connector=UCMConnector` option becomes available through the
configuration interface.
- When selected, vllm-ascend workers may initialize UCM and offload
KV-cache blocks externally.
- No default behaviors are changed. Users must explicitly enable this
connector.
This PR does **not** modify:
- existing APIs,
- default execution paths,
- model runner behavior,
- user workflow unless `UCMConnector` is configured.
---
### How was this patch tested?
---
### Prefix Caching Benchmark
We provide preliminary measurements for TTFT (ms) under VLLM benchmark.
Tests run on 2 * Ascend 910B3, vllm-ascend 0.11.0, Tensor Parallel size
2, with UCM (Localdisk) enabled.
- vLLM version: v0.12.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/ad32e3e19ccf0526cb6744a5fed09a138a5fb2f9
Signed-off-by: UnifiedCacheManager <unifiedcachem@163.com>
### What this PR does / why we need it?
This Pull Request removes the @pytest.mark.skip decorators from
test_mtp1_correctness_piecewise_graph and
test_mtp2_correctness_piecewise_graph.
These tests were temporarily skipped because of an issue with the MTP
ACL Graph (as per the original TODO comment). Since the relevant
bug/issue has been resolved, these tests are now re-enabled to ensure
full correctness coverage for MTP functionality.
- vLLM version: v0.12.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/ad32e3e19ccf0526cb6744a5fed09a138a5fb2f9
Signed-off-by: MrZ20 <2609716663@qq.com>
Co-authored-by: Mengqing Cao <cmq0113@163.com>
### What this PR does / why we need it?
In the current KV Pool scenario for models like MLA and GQA, where
different TP ranks generate identical KV caches, the system is designed
to store only a single copy. The previous approach allowed each card to
query storage requirements dynamically, but inconsistent query results
across cards led to incorrect storage. To fix this, the new solution
pre-allocates storage responsibilities; each card now simply stores its
pre-assigned blocks, bypassing the inconsistent query step and ensuring
data correctness.
- vLLM version: v0.12.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/ad32e3e19ccf0526cb6744a5fed09a138a5fb2f9
---------
Signed-off-by: fems14 <1804143737@qq.com>
### What this PR does / why we need it?
This PR adds back pa in scenarios of small batch sizes due to
performance consideration. Will remove pa once fia performs better than
pa in all scenarios.
### Does this PR introduce _any_ user-facing change?
No
### How was this patch tested?
CI passed with existing test.
- vLLM version: v0.12.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/ad32e3e19ccf0526cb6744a5fed09a138a5fb2f9
---------
Signed-off-by: whx-sjtu <2952154980@qq.com>
Co-authored-by: weijinqian0 <1184188277@qq.com>
### What this PR does / why we need it?
Modify quantization tutorials to correct a few mistakes:
Qwen3-32B-W4A4.md and Qwen3-8B-W4A8.md
Qwen3-8B-W4A8: need to set one idle npu card.
Qwen3-32B-W4A4: need to set two idle npu cards for the flatquant
training and modify the calib_file path which does not match the
ModeSlim version.
### Does this PR introduce _any_ user-facing change?
N/A
### How was this patch tested?
- vLLM version: v0.12.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/ad32e3e19ccf0526cb6744a5fed09a138a5fb2f9
Signed-off-by: IncSec <1790766300@qq.com>
### What this PR does / why we need it?
Fix the incorrect use of python's sum function on PyTorch tensors.
1. Using Python's sum() function on a tensor self.num_pcp_pads resulted
in 6ms execution time
Optimization: replacing with PyTorch's torch.sum() reduced execution
time to 474µs
2. scheduler_output.scheduled_spec_decode_tokens undergoes repeated loop
processing even when speculative decoding is not used
Optimization: added conditional logic to skip processing loops when
speculative decoding is disabled, eliminating unnecessary computational
overhead.
- vLLM version: 86e178f7c4d8c3b0eaf3c8e3f810a83f63b90e24
- vLLM main:
https://github.com/vllm-project/vllm/commit/86e178f7c4d8c3b0eaf3c8e3f810a83f63b90e24
Signed-off-by: wangx700 <wangxin700@huawei.com>
Co-authored-by: weijinqian0 <1184188277@qq.com>
### What this PR does / why we need it?
Delete accuracy tests for models that are no longer retained:
- Meta-Llama-3.1-8B-Instruct
- llava-1.5-7b-hf
- InternVL2-8B.yaml
- InternVL2_5-8B.yaml
- InternVL3-8B.yaml
Add accuracy tests for the new models:
- Llama-3.2-3B-Instruct
- llava-onevision-qwen2-0.5b-ov-hf
- Qwen3-VL-30B-A3B-Instruct
- vLLM version: v0.12.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/ad32e3e19ccf0526cb6744a5fed09a138a5fb2f9
---------
Signed-off-by: MrZ20 <2609716663@qq.com>
### What this PR does / why we need it?
- Fix a premature `return` in `moe_init_routing_quant_v2.cpp` so the
routing kernel completes correctly instead of exiting early in certain
paths.
- Switch `FusedAlltoAllCommImpl` to use the MC2-based token dispatcher
and prepare/finalize routines, aligning MoE communication with the MC2
algorithm optimized for Ascend devices.
- Add a temporary override in `MtpProposer` to map `FUSED_ALLTOALL` back
to `ALLTOALL` until the MoE communication type selection logic is fully
finalized, avoiding incorrect behavior in dummy-run flows.
- Simplify the MoE communication selection for Ascend 910-93 in
`NPUModelRunner` by removing the EP-size guard on `FUSED_ALLTOALL`,
which fixes failures in multi-node / larger-EP configurations while
keeping MC2 routing under the configured token capacity.
- vLLM version: v0.12.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/ad32e3e19ccf0526cb6744a5fed09a138a5fb2f9
Signed-off-by: mojave2 <chenchen145@huawei.com>
### What this PR does / why we need it?
This PR provides an introduction to the Qwen3-Next model, details on the
features supported by the model in the current version, the model
deployment process, as well as methods for performance testing and
accuracy testing.
With this document, the deployment and testing of the Qwen3-Next model
can be implemented more easily.
- vLLM version: v0.12.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/ad32e3e19ccf0526cb6744a5fed09a138a5fb2f9
---------
Signed-off-by: ming1212 <2717180080@qq.com>
Signed-off-by: ming1212 <104972349+ming1212@users.noreply.github.com>
Signed-off-by: Mengqing Cao <cmq0113@163.com>
Co-authored-by: Mengqing Cao <cmq0113@163.com>