sherie f86596a66c allgather use fusedop. (#2689) ...

### What this PR does / why we need it?
Use 'npu_moe_init_routing_v2' &'npu_moe_token_unpermute' repalce
'npu_moe_init_routing' &‘npu_moe_compute_expert_tokens’&
'npu_moe_finalize_routing' to optimize performance
### Does this PR introduce _any_ user-facing change?
| branch| tps| TTFT |TPOT |
| --- | --- | --- |--- |
|main  |733.98  | 280.05 |34.30 |
|main+fusedop  | 740.33 | 273.34 |33.99 |
### How was this patch tested?


- vLLM version: v0.10.1.1
- vLLM main:
6997a25ac6

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

2025-09-04 11:56:29 +08:00

..

attention

[BugFix][MLA] Fix attn_mask bug for ring mla (#2704)

2025-09-04 10:22:46 +08:00

core

[CI] fix UT error. (#2644)

2025-08-30 12:04:01 +08:00

device_allocator

add ut for device allocator/camem and mutistream/layers (#2037)

2025-07-31 19:17:27 +08:00

distributed

[Feat]: Add custom lmhead tensor model parallel (#2309)

2025-08-29 11:41:21 +08:00

fake_weight

[CI] Add unit test framework (#1201)

2025-06-16 18:32:28 +08:00

kv_connector

[P/D]mooncake_connector adapted to 0.10.1 (#2664)

2025-09-04 08:22:10 +08:00

models

[Bugfix] Fix qwen2.5-vl-without-padding (#2623)

2025-09-03 14:38:55 +08:00

multistream

add ut for device allocator/camem and mutistream/layers (#2037)

2025-07-31 19:17:27 +08:00

ops

allgather use fusedop. (#2689)

2025-09-04 11:56:29 +08:00

patch/worker/patch_common

[Misc] Remove redundant imported envs, using envs_ascend instead (#2193)

2025-08-14 09:33:39 +08:00

quantization

[main] Fuse GroupedMatmul, Swiglu and DynamicQuant in W8A8_DYNAMIC quantized MoE layers (#2275)

2025-09-04 11:37:32 +08:00

sample

[Perf][MTP] Optimize reject sampler in greedy situation. (#2137)

2025-08-11 17:37:49 +08:00

torchair

[1/N][Refactor][Quantization] remove redundant quantizer class (#2680)

2025-09-04 11:35:14 +08:00

worker

[Main][Feat]Set the Profiler parameters through environment variables consistent with vLLM (#2608)

2025-09-03 10:58:08 +08:00

__init__.py

[2/4][Refactor] Refactor torchair utils (#1892)

2025-07-21 19:43:30 +08:00

base.py

[Feature]: implement the fusion of allreduce and matmul in prefill phase when tp is enabled (#1926)

2025-07-28 15:13:37 +08:00

conftest.py

[1/N][CustomOp] Register activation customop instead of overwrite forward_oot (#1841)

2025-07-18 23:07:14 +08:00

test_ascend_config.py

[bugfix] fix torchair runtime error caused by configuration mismtaches and file missing (#2532)

2025-09-03 17:56:12 +08:00

test_envs.py

[Misc] Remove redundant imported envs, using envs_ascend instead (#2193)

2025-08-14 09:33:39 +08:00

test_platform.py

[Feat] allow using aclgraph in ray backend (#2589)

2025-09-04 11:45:56 +08:00

test_utils.py

[Fix] fix resources limit error when apply speculative decoding and aclgraph (#2472)

2025-09-04 11:50:43 +08:00