Skip to content

Conversation

@fxyfxy777
Copy link
Contributor

  1. cublas的输入是int,我们在外面使用更大的类型也没有用
  2. n是矩阵的边长,边长不可能超过int(哪怕n^2也不可能超过int,不然数值不稳定解不出来)
    可参考[PHI][CINN] Align linalg.solve kernel with torch PaddlePaddle/Paddle#72608

@paddle-bot
Copy link

paddle-bot bot commented Jul 29, 2025

Thanks for your contribution!

Comment on lines 1801 to 1802
paddle.linalg.solve(Tensor([10, 10],"float32"), Tensor([10, 228170138],"float32"), )
paddle.linalg.solve(Tensor([10, 10],"float32"), Tensor([228170138, 10],"float32"), left=False, )
Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

考虑在 paddle 中加入显式报错🧐?然后再添加到 paddle_error_dismiss

参考:PaddlePaddle/Paddle#73194

Copy link
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

ok,我改下

@cangtianhuang cangtianhuang self-assigned this Jul 29, 2025
@wanghuancoder wanghuancoder merged commit 79f2490 into PFCCLab:main Jul 30, 2025
@fxyfxy777 fxyfxy777 deleted the fix_linalg_solve branch July 30, 2025 08:04
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants