(this PR is just for my personal self-learning) [PaddlePaddle Hackathon 2 No.22] add paddle.index_add to Paddle

[OccupyMars2025]

PR types

New features

PR changes

APIs

Describe

【PaddlePaddle Hackathon 第二期】任务总览
task 22: add paddle.index_add to Paddle
#40311

RFC： PaddlePaddle/community#127

This PR is gotten by modifying this PR: #42454
(with least modification)

I stop working on this PR at 2022/5/17 18:00 . I will continue to work on this PR after 2022/6/10 when I have time.

[paddle-bot-old]

你的PR提交成功，感谢你对开源项目的贡献!
请关注后续CI自动化测试结果，详情请参考Paddle-CI手册。
Your PR has been submitted. Thanks for your contribution!
Please wait for the result of CI firstly. See Paddle CI Manual for details.

[OccupyMars2025]

还是有其他算子的单侧报错了

[OccupyMars2025]

Caution: add "x = np.copy(x)" at the first line of the function that use numpy to simulate the calculation of paddle op in unit test file. Because if you don't use the copy of input at first , then the calculation in the function may change the elements of numpy.ndarray in place, so after calling this function, the input numpy.ndarray is modified to be equal to the output numpy.ndarray

TODO: #42454
in the above PR, there seems to be no such 'copy' thing, how it passes the unit test ???
remain to be studied...

[OccupyMars2025]

再次触发CI后，发现报错信息完全与我的代码无关，是其他算子的单测有问题
解决方法： Merge PaddlePaddle:develop 到我的分支，重新触发CI

[OccupyMars2025]

已经Merge PaddlePaddle:develop 到我的分支，重新触发CI，发现报错信息仍然完全与我的代码无关，是其他算子的单测有问题

[OccupyMars2025]

重新触发CI，发现报错信息仍然完全与我的代码无关，是其他算子的单测有问题
解决办法：提交其他PR，看是否是所有的PR都有这个问题

[OccupyMars2025]

#42804
看看这个PR是否会报同样的错误, 显示没有报同样的错误，说明是你修改了与任务无关的代码

[OccupyMars2025]

你确实修改了与任务无关的代码

[OccupyMars2025]

网上搜索发现：该报错似乎都与PaddlePaddle 训练或者infer相关，似乎是CUDA, cuDNN 版本问题？？？

单测文件中只留下第一个test case, 其他的全部注释掉，重新提交，又可以了， 没有上面的"Segmentation fault"报错问题了

[OccupyMars2025]

感谢Paddle 专家的指导意见，我将暂停本PR的开发工作，在以后有时间时再继续开发

[paddle-bot-old]

Sorry to inform you that 24d3827's CIs have passed for more than 7 days. To prevent PR conflicts, you need to re-run all CIs manually.

[paddle-bot]

Since you haven't replied for more than a year, we have closed this issue/pr.
If the problem is not solved or there is a follow-up one, please reopen it at any time and we will continue to follow up.
由于您超过一年未回复，我们将关闭这个issue/pr。
若问题未解决或有后续问题，请随时重新打开，我们会继续跟进。

[paddle-bot]

很抱歉，经过我们的反复讨论，你的PR暂未达到合入标准，请阅读飞桨原生算子开发规范，你可以重新提交新的PR，我们先将此PR关闭，感谢你的贡献。
Sorry to inform you that through our discussion, your PR fails to meet the merging standard (Reference: Paddle Custom Operator Design Doc). You can also submit an new one. Thank you.