已合并
docs(gmm): 明确MXFP8量化中存储block与group size的概念区分 #9736
zhoushaolong创建于 9 天前
docs(gmm): 明确MXFP8量化中存储block与group size的概念区分 #9736
已合并
Pull Request已成功合入, 合并人@CANN-robot
(感谢 zhoushaolong 的贡献)9 天前 创建了 pull request,commit a6fb846f
atomgit-bot
9 天前 评论:
9 天前 评论:
9 天前 添加了label:cann-cla/yes
CANN-robot
9 天前 评论:
9 天前 评论:
CANN-robot
9 天前 评论:
9 天前 评论:
Thanks for your pull-request.
The full list of commands accepted by me can be found at here。
You can get sig-info at here
PR Approval Progress
✅ Congratulations! All modules have met the lgtm and approve requirements.
Module Approval Details
| module | lgtm status | approve status |
|---|---|---|
| */*/docs/acl*.md | ✅ 陈娇, 蒋礼锐 (2/2) | ✅ 陈娇 (1/1) |
💡 Tip:
- Committer can comment
/approveor/lgtm- Commenting
/approveimplies both code review (lgtm) and intent to merge (approve)


9 天前 将monologue815,jiang-lirui,Allan_Yu,gitee-duhuiping,claire8090,liudan12,gitcode-chenjiao,Ww7012,zhouwenfang,mabing1118,wuyao51511,caiwenwen设为评审人
9 天前 将monologue815,jiang-lirui,Allan_Yu,gitee-duhuiping,claire8090,liudan12,gitcode-chenjiao,Ww7012,zhouwenfang,mabing1118,wuyao51511,caiwenwen设为评审人
9 天前 将gitee-duhuiping,claire8090,gitcode-chenjiao,Ww7012,zhouwenfang,wuyao51511,caiwenwen设为审查人
9 天前 将gitee-duhuiping,claire8090,gitcode-chenjiao,Ww7012,zhouwenfang,wuyao51511,caiwenwen设为审查人
zhoushaolong
9 天前 评论:
9 天前 评论:
compile


9 天前 update merge request[project id: 7673863, iid: 9736, commit_id: d3236015aea9bf4251159d50efab3caed2d430c0] virtual merging success
9 天前 update merge request[project id: 7673863, iid: 9736, commit_id: d3236015aea9bf4251159d50efab3caed2d430c0] virtual merging success
9 天前 添加了label:ci-pipeline-running
CANN-robot
9 天前 评论:
9 天前 评论:
流水线任务触发成功
任务链接 [9d75c8f4e9d04c2fa9cc6617524329fe][流水线指导]
| 任务名称 | 状态 | 日志 | 下载链接 |
|---|---|---|---|
| Compile_Ascend_X86 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_ARM | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_X86_experimental | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_ARM_experimental | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Pre | ✅ SUCCESS | >>>>> | |
| pre_comment | ✅ SUCCESS | >>>>> | |
| Compile_Ascend_X86_ubuntu24 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_ARM_ubuntu24 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_X86_experimental_ubuntu24 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_ARM_experimental_ubuntu24 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_X86_monitor_910b | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_X86_monitor_910c | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_X86_monitor_950 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_X86_experimental_950 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_ARM_experimental_950 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_X86_experimental_950_ubuntu24 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_ARM_experimental_950_ubuntu24 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_classify | ✅ SUCCESS | >>>>> | |
| Compile_Ascend_X86_mobile_station_ubuntu24 | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_X86_mobile_station | ✅ SUCCESS | >>>>> | >>>>> |
| Compile_Ascend_X86_mobile_station_9030_ubuntu24 | ✅ SUCCESS | >>>>> | >>>>> |
| UT_Test | ✅ SUCCESS | ||
| PreSmoke_A900 | ✅ SUCCESS | >>>>> | |
| API_Check | ✅ SUCCESS | >>>>> | |
| PreSmoke_ATK_Test_A2 | ✅ SUCCESS | >>>>> | |
| UT_Test_harmony-infer-chs-transformer | ✅ SUCCESS | ||
| UT_Test_harmony-infer-cpro-transformer | ✅ SUCCESS | ||
| UT_Test_report_lcov | ⚠️ WARNING | >>>>> |
[2026-08-07 11:53:16] CI执行结束


CANN-robot
9 天前 评论:
9 天前 评论:
流水线任务触发成功
任务链接 [9e435d1313ad44b884d50a92da4acf35][流水线指导]
| 任务名称 | 状态 | 日志 | 下载链接 |
|---|---|---|---|
| codecheck | ✅ SUCCESS | >>>>> | |
| SCA | ✅ SUCCESS | >>>>> | |
| antipoison | ✅ SUCCESS | >>>>> | |
| codecheck_style | ✅ SUCCESS | >>>>> | |
| StaticCheck_codespell | ✅ SUCCESS | ||
| StaticCheck_link_validity | ✅ SUCCESS | ||
| StaticCheck_resource_existence | ✅ SUCCESS | ||
| StaticCheck_tag_closed | ✅ SUCCESS | ||
| StaticCheck_markdownlint | ✅ SUCCESS | ||
| codecheck_checkpr | ✅ SUCCESS | ||
| codecheck_precommit | ✅ SUCCESS | >>>>> |
[2026-08-07 11:34:11] CI执行结束


9 天前 修改标题为 “docs(gmm): 明确MXFP8量化中存储block与group size的概念区分”,原标题为“docs: clarify MXFP8 block and group sizes”
9 天前 修改标题为 “docs(gmm): 明确MXFP8量化中存储block与group size的概念区分”,原标题为“docs: clarify MXFP8 block and group sizes”
此处折叠了5条事件消息 查看更多
9 天前 添加了label:ci-pipeline-passed
9 天前 添加了label:approved
jiang-lirui
6 天前 评论:
6 天前 评论:
/approve
/lgtm


6 天前 添加了label:lgtm
6 天前 关闭了关联的issue
6 天前 合入了pull request
描述
明确 aclnnGroupedMatmulActivationQuantWeightNz 算子文档中 MXFP8 量化的"存储block大小"与"MX量化group大小"两个概念,消除原表述统一使用 blocksize 造成的混淆。
改动原因
原文档统一用 blocksize 描述 MXFP8 量化的分组大小,未区分两个不同概念:
两者数值不同,混用会导致读者对量化分组逻辑(尤其 scaleAlg=0 与 scaleAlg=1 两种场景)理解偏差。
改动方法
关联的Issue
https://gitcode.com/cann/ops-transformer/issues/4170
测试
纯文档改动,不涉及代码逻辑变更,无需测试。
文档更新
更新了 gmm/grouped_matmul_activation_quant/docs/aclnnGroupedMatmulActivationQuantWeightNz.md:明确 MXFP8 存储block(64)与量化group(32)的概念区分,修正 scaleAlg=0/1 两种场景的分组参数表述。
类型标签