Logo
Explore Help
Register Sign In
EngineX/xc-llm-ascend
Watch 3
Star 0 Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity
Files
63adbedb7aeabdabd7c90dd9cea2f4e90eb320e0
xc-llm-ascend/tests/ut/attention
T
History
Qiu 749e24f81e [bugfix] align max_num_batched_tokens with tp*pcp when using FLASHCOMM1 (#6000)
### What this PR does / why we need it?
Align max_num_batched_tokens with tp*pcp when using FLASHCOMM1 to avoid
assert error in `NPUModelRunner._dummy_run`.

- vLLM version: v0.13.0
- vLLM main:
https://github.com/vllm-project/vllm/commit/2c24bc6996cb165fce92f780b388a5e39b3f4060

---------

Signed-off-by: QiuChunshuo <qiuchunshuo@huawei.com>
2026-01-23 14:19:49 +08:00
..
test_attention_cp.py
[Main2Main] Upgrade vllm commit to 0109 (#5752)
2026-01-13 19:14:43 +08:00
test_attention_mask.py
[Refactor] 2/N Unify all mask generation methods and cache mask (#4779)
2025-12-09 18:51:00 +08:00
test_attention_v1.py
[Main2Main] Upgrade vllm commit to 0113 (#5839)
2026-01-15 09:48:53 +08:00
test_mla_cp.py
[bugfix] align max_num_batched_tokens with tp*pcp when using FLASHCOMM1 (#6000)
2026-01-23 14:19:49 +08:00
test_mla_v1.py
[bugfix] align max_num_batched_tokens with tp*pcp when using FLASHCOMM1 (#6000)
2026-01-23 14:19:49 +08:00
test_sfa_v1.py
[bugfix] align max_num_batched_tokens with tp*pcp when using FLASHCOMM1 (#6000)
2026-01-23 14:19:49 +08:00
utils.py
[refactor](UT,PCP,DCP) refactor pcp&dcp patches in UTs (#5505)
2026-01-05 09:05:45 +08:00
Powered by Gitea Version: 28.4.0 Page: 45ms Template: 3ms
Auto
English
Bahasa Indonesia Deutsch English Español Français Gaeilge Italiano Latviešu Magyar nyelv Nederlands Polski Português de Portugal Português do Brasil Suomi Svenska Türkçe Čeština Ελληνικά Български Русский Українська فارسی മലയാളം 日本語 简体中文 繁體中文(台灣) 繁體中文(香港) 한국어
Licenses API