Song Mingyang 18b90b501d [kernel] add AscendC op: lightning_indexer and sparse_flash_attention (#4625) ...

### What this PR does / why we need it?
Provide high-performance AscendC operators lightning_indexer and
sparse_flash_attention to boost the execution performance of the
DeepSeek v3.2 model. Meanwhile, adapt the two AscendC operators to
vllm-ascend framework.

### Does this PR introduce _any_ user-facing change?
No (only underlying operator optimizations, with no user-facing changes)

### How was this patch tested?

- vLLM version: v0.11.2
- vLLM main: https://github.com/vllm-project/vllm/commit/v0.11.2

Signed-off-by: MingYang119 <songmingyang@huawei.com>

2025-12-03 09:53:10 +08:00

..

_cann_ops_custom

[Kernel] add custom op GmmSwigluQuantWeightNzTensorList (#3804)

2025-11-28 18:06:39 +08:00

attention

[kernel] add AscendC op: lightning_indexer and sparse_flash_attention (#4625)

2025-12-03 09:53:10 +08:00

compilation

upgrade to vllm 0.11.2 (#4400)

2025-11-26 11:48:58 +08:00

core

upgrade vLLM to main (#4608)

2025-12-02 22:10:52 +08:00

device_allocator

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

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

distributed

upgrade vLLM to main (#4608)

2025-12-02 22:10:52 +08:00

eplb

[EPLB][Ops] Integerate grouped_matmul_swiglu_quant_weight_nz_tensor_list operator into dynamic EPLB (#4216)

2025-11-30 22:52:05 +08:00

kv_offload

upgrade vLLM to main (#4608)

2025-12-02 22:10:52 +08:00

lora

[refact] unified soc_version code (#4359)

2025-11-26 14:28:55 +08:00

model_loader

Drop 0.11.0 support (#4377)

2025-11-24 17:08:20 +08:00

ops

upgrade vLLM to main (#4608)

2025-12-02 22:10:52 +08:00

patch

upgrade vLLM to main (#4608)

2025-12-02 22:10:52 +08:00

quantization

[main][bugfix] bugfix for qwen3 moe quantization (#4599)

2025-12-01 23:48:57 +08:00

sample

[performance] Enhance performance after enabling min_p (#4529)

2025-12-02 20:35:51 +08:00

spec_decode

upgrade vLLM to main (#4608)

2025-12-02 22:10:52 +08:00

torchair

upgrade vLLM to main (#4608)

2025-12-02 22:10:52 +08:00

worker

[kernel] add AscendC op: lightning_indexer and sparse_flash_attention (#4625)

2025-12-03 09:53:10 +08:00

__init__.py

clean up model module (#4611)

2025-12-02 17:35:47 +08:00

ascend_config.py

[Feat] shared expert dp for deepseek_mtp (#3811)

2025-12-01 20:44:11 +08:00

ascend_forward_context.py

[Refactor] remove moe type of multicast. (#4224)

2025-11-24 17:32:37 +08:00

cpu_binding.py

[main] support cpu binding (#3546)

2025-10-21 09:17:03 +08:00

envs.py

[refact] unified soc_version code (#4359)

2025-11-26 14:28:55 +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

upgrade vLLM to main (#4608)

2025-12-02 22:10:52 +08:00

profiling_config.py

Revert "drop ascend scheduler" (#4580)

2025-11-29 22:20:48 +08:00

utils.py

upgrade vLLM to main (#4608)

2025-12-02 22:10:52 +08:00