LI SHENGYONG 0331f16a50 [EPLB] Reduce the memory used for heat aggregation (#6729) ...

### What this PR does / why we need it?
If dist.all_gather is used directly, 2 x HCCL_BUFFSIZE memory will be
consumed, but the actual memory required for hotspot aggregation is less
than 1 MB. Therefore, a separate small communication domain is created
for it.

### Does this PR introduce _any_ user-facing change?

### How was this patch tested?
Original：

![1](https://github.com/user-attachments/assets/8880b461-c26f-497c-9a05-2ca60cc46aa4)
Current：

![2](https://github.com/user-attachments/assets/c9da32b5-9200-4fa2-aff9-d8c4978ac602)


- vLLM version: v0.15.0
- vLLM main:
9562912cea

Signed-off-by: shenchuxiaofugui <1311027364@qq.com>

2026-02-24 18:02:24 +08:00

..

_310p

[Feat] 310p supports PrefillCacheHit State (#6756)

2026-02-24 16:48:05 +08:00

_cann_ops_custom

[Kernel] add custom op GmmSwigluQuantWeightNzTensorList (#3804)

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

attention

[P/D][PCP] mooncake layerwise support pcp function (#6627)

2026-02-12 11:02:25 +08:00

compilation

[CI]Fixed the spell check function in typos.toml (#6753)

2026-02-14 11:57:26 +08:00

core

[CI]Fixed the spell check function in typos.toml (#6753)

2026-02-14 11:57:26 +08:00

device

[Lint]Style: Convert vllm-ascend/ to ruff format(Batch #2) (#5977)

2026-01-19 08:59:46 +08:00

device_allocator

[Lint]Style: Convert vllm-ascend/ to ruff format(Batch #2) (#5977)

2026-01-19 08:59:46 +08:00

distributed

[EPLB] Reduce the memory used for heat aggregation (#6729)

2026-02-24 18:02:24 +08:00

eplb

[EPLB] Reduce the memory used for heat aggregation (#6729)

2026-02-24 18:02:24 +08:00

kv_offload

[main2main] upgrade vllm main 0202 (#6560)

2026-02-05 19:31:17 +08:00

lora

[Lint]Style: Convert vllm-ascend/ to ruff format(Batch #5) (#5996)

2026-01-24 22:45:38 +08:00

model_loader

[Lint]Style: Convert vllm-ascend/ to ruff format(Batch #6) (#6001)

2026-01-24 22:08:33 +08:00

ops

[Bugfix][DispatchFFNCombine] resolve vec error caused by unaligned UB access (#6707)

2026-02-14 10:32:50 +08:00

patch

[Feature] adapt to uva buffer and main2main (#6657)

2026-02-12 10:36:31 +08:00

quantization

[Model] GLM5 adaptation (#6642)

2026-02-11 22:22:22 +08:00

sample

[Bugfix] Update target probs to target logits in rejection sample (#6685)

2026-02-11 21:31:40 +08:00

spec_decode

[CI]Fixed the spell check function in typos.toml (#6753)

2026-02-14 11:57:26 +08:00

worker

[EPLB][Bugfix] Bugfix for ineffective dynamic eplb (#6653)

2026-02-24 14:43:04 +08:00

xlite

[Lint]Style: Convert vllm-ascend/ to ruff format(Batch #10) (#6173)

2026-02-06 15:35:06 +08:00

__init__.py

[Lint]Style: Convert vllm-ascend/compilation to ruff format (#5912)

2026-01-16 20:57:46 +08:00

ascend_config.py

[CI]Fixed the spell check function in typos.toml (#6753)

2026-02-14 11:57:26 +08:00

ascend_forward_context.py

[Main][Ops] Make triton rope support index_selecting from cos_sin_cache (#5450)

2026-02-11 21:20:53 +08:00

batch_invariant.py

implement batch invariant with ascendc (#6590)

2026-02-10 14:15:26 +08:00

cpu_binding.py

[Lint]Style: Convert vllm-ascend/compilation to ruff format (#5912)

2026-01-16 20:57:46 +08:00

envs.py

[MISC] Clean up useless env USE_OPTIMIZED_MODEL (#6618)

2026-02-09 15:38:58 +08:00

flash_common3_context.py

[Lint]Style: Convert vllm-ascend/compilation to ruff format (#5912)

2026-01-16 20:57:46 +08:00

meta_registration.py

[Ops][Refactor] Remove custom rotary_embedding operator (#6523)

2026-02-07 09:24:05 +08:00

platform.py

[CI]Fixed the spell check function in typos.toml (#6753)

2026-02-14 11:57:26 +08:00

profiling_config.py

[Core][Misc] Clean up ProfileExecuteDuration (#6461)

2026-02-01 20:06:01 +08:00

utils.py

[EPLB] Reduce the memory used for heat aggregation (#6729)

2026-02-24 18:02:24 +08:00