XiaoxinWang 9eb62935b8 fix pagedattention to support fullgraph. (#3436) ...

### What this PR does / why we need it?
Calculate in advance the workspace memory size needed for the
PagedAttention operator to avoid deadlocks during resource cleanup. This
PR requires torch_npu version 0920 or newer.
### Does this PR introduce _any_ user-facing change?

### 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: wangxiaoxin-sherie <wangxiaoxin7@huawei.com>
Co-authored-by: wangxiaoxin-sherie <wangxiaoxin7@huawei.com>

2025-10-14 16:10:09 +08:00

..

attention

fix pagedattention to support fullgraph. (#3436)

2025-10-14 16:10:09 +08:00

compilation

fix pagedattention to support fullgraph. (#3436)

2025-10-14 16:10:09 +08:00

core

[BugFix] Fix ascend scheduler assert error (#3191)

2025-09-28 18:22:08 +08:00

device_allocator

[Misc]Clean up useless import from vllm (#2049)

2025-07-28 16:01:59 +08:00

distributed

[Feature] mooncake connector support GQA transport (#2947)

2025-10-13 15:48:37 +08:00

eplb

Bugfix: Expose the user policy type interface (#3336)

2025-10-11 16:28:57 +08:00

lora

[Bugfix][LoRA] Fix forward error and shape mismatch when using LoRA (#3153)

2025-09-28 17:30:50 +08:00

models

[MoE] [Refactor] Combine common_fused_moe and fused_moe (#3176)

2025-10-09 14:12:46 +08:00

multistream

[Quickfix] update CachedRequestState as NewRequestData changed (#2367)

2025-08-15 07:35:27 +08:00

ops

[Feature] optimize sp & qwen3 next support sp. (#3225)

2025-10-13 23:02:12 +08:00

patch

[feat] support customized and separated hccl_buffer_size for process group initialization (#3073)

2025-10-11 15:55:22 +08:00

quantization

[Feature] Add W4A4 Flat Quantization support (#3427)

2025-10-13 23:20:16 +08:00

sample

Drop 0.10.2 (#3284)

2025-10-09 10:28:38 +08:00

spec_decode

bugfix for mtp (#3300)

2025-10-09 19:22:46 +08:00

torchair

[Feature] optimize sp & qwen3 next support sp. (#3225)

2025-10-13 23:02:12 +08:00

worker

[Fix] Fix mc2_tokens_capacity-related issues (#3411)

2025-10-14 10:56:12 +08:00

__init__.py

【bugfix】fix connector register failed (#3335)

2025-10-09 21:09:54 +08:00

ascend_config.py

Bugfix: Expose the user policy type interface (#3336)

2025-10-11 16:28:57 +08:00

ascend_forward_context.py

Revert PTA upgrade PR (#3352)

2025-10-10 14:09:53 +08:00

envs.py

Add DeepSeek V3.2 support (#3270)

2025-09-30 03:25:58 +08:00

meta_registration.py

Fix the bugs about operator registration by PyTorch Dispatcher (#2786)

2025-09-13 11:58:52 +08:00

platform.py

[Feat][Graph]Support FULL_DECEDE_ONLY mode for MLA models (#3125)

2025-10-10 16:31:20 +08:00

utils.py

[Feat] enable hierarchical communication for mc2 ops on A2 (#3015)

2025-10-13 16:13:17 +08:00