-
Notifications
You must be signed in to change notification settings - Fork 18
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Fix gpu unit test. #545
Fix gpu unit test. #545
Conversation
hunghaoti
commented
Dec 27, 2024
- Fix Lanczos Gnd unit test cause from putblock using different devices.
- Skip ExpM for gpu test since Eig still not implemted, which is needed to do the exponential.
- cuda svd cannot give the correct answer if almost elements are zeros. So we skip Svd U1_zeros tests.
- The answer of Svd_truncate3 is not correct also cause from the above issue. So I delete it.
1. Fix Lanczos Gnd unit test cause from putblock using different devices. 2. Skip ExpM for gpu test since Eig still not implemted, which is needed to do the exponential. 3. cuda svd cannot give the correct answer if almost elements are zeros. So we skip Svd U1_zeros tests. 4. The answer of Svd_truncate3 is not correct also cause from the above issue. So I delete it.
Codecov ReportAll modified and coverable lines are covered by tests ✅
Additional details and impacted files@@ Coverage Diff @@
## dev-master #545 +/- ##
===========================================
Coverage 23.58% 23.58%
===========================================
Files 211 211
Lines 44926 44926
Branches 14018 14018
===========================================
Hits 10595 10595
Misses 32664 32664
Partials 1667 1667 ☔ View full report in Codecov by Sentry. |
Just leave a reference. The third problem was addressed in #505. |
There was a problem hiding this comment.
Choose a reason for hiding this comment
The reason will be displayed to describe this comment to others. Learn more.
Does the first problem imply there is a bug in the function to put blocks while the device is the GPU?
Feel free to close the conversations to merge the PR. |
The first problem cause since we cannot put a Tensor in GPU device to the UniTensor in CPU device. So we just create Tensors in CPU and put it in UniTensor in CPU. And then transfer to GPU finally. |