Feng Liu 49838d4bec [Perf] vectorize PCP/DCP loops in attention_cp.py (#4944) ...

### What this PR does / why we need it?
- Add explicit .contiguous() after permute/view to ensure mem-friendly
layout
- Replace nested PCP/DCP Python loops with fully vectorized tensor
operations

- vLLM version: v0.12.0
- vLLM main:
ad32e3e19c

---------

Signed-off-by: F.Liu <liufeng248@huawei.com>
Co-authored-by: F.Liu <liufeng248@huawei.com>

2025-12-22 11:06:19 +08:00

..

_cann_ops_custom

[Kernel] add custom op GmmSwigluQuantWeightNzTensorList (#3804)

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

attention

[Perf] vectorize PCP/DCP loops in attention_cp.py (#4944)

2025-12-22 11:06:19 +08:00

compilation

[Refactor] remove some metadata variables in attention_v1. (#5160)

2025-12-19 14:57:09 +08:00

core

[CI] fix lint (#5216)

2025-12-20 17:03:25 +08:00

device_allocator

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

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

distributed

fix vl pd smoke error (#5103)

2025-12-18 22:20:45 +08:00

eplb

[Misc] Upgrade vllm hash to 12_14 (#5000)

2025-12-15 19:54:23 +08:00

kv_offload

upgrade vLLM to main (#4608)

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

lora

[BugFix]Fix precision issue for LoRA feature (#4141)

2025-12-19 14:22:06 +08:00

model_loader

[CI] speed up ut (#4901)

2025-12-11 18:45:43 +08:00

ops

[Feature]Use DispatchGmmCombineDecode operator to replace MC2(Optional) (#5040)

2025-12-21 15:23:59 +08:00

patch

Drop 0.12.0 support (#5146)

2025-12-20 09:38:53 +08:00

quantization

[Feature]Use DispatchGmmCombineDecode operator to replace MC2(Optional) (#5040)

2025-12-21 15:23:59 +08:00

sample

[Performance] Add async exponential while model executing (#4501)

2025-12-20 21:23:21 +08:00

spec_decode

[Refactor] remove some metadata variables in attention_v1. (#5160)

2025-12-19 14:57:09 +08:00

worker

[Performance] Add async exponential while model executing (#4501)

2025-12-20 21:23:21 +08:00

xlite

[Refactor] remove some metadata variables in attention_v1. (#5160)

2025-12-19 14:57:09 +08:00

__init__.py

clean up model module (#4611)

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

ascend_config.py

[Performance] Add async exponential while model executing (#4501)

2025-12-20 21:23:21 +08:00

ascend_forward_context.py

[Feature]Use DispatchGmmCombineDecode operator to replace MC2(Optional) (#5040)

2025-12-21 15:23:59 +08:00

cpu_binding.py

[main] support cpu binding (#3546)

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

envs.py

[Feature]Use DispatchGmmCombineDecode operator to replace MC2(Optional) (#5040)

2025-12-21 15:23:59 +08:00

flash_common3_context.py

[Perf]enable prefill flashcommon3 (#4065)

2025-12-14 09:34:13 +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

Drop 0.12.0 support (#5146)

2025-12-20 09:38:53 +08:00

profiling_config.py

Drop ascend scheduler (#4623)

2025-12-05 09:03:45 +08:00

utils.py

[BugFix]Fix wrong _cos, _sin instantiation (#5154)

2025-12-20 22:52:50 +08:00