Skip to content
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

Convolution self test fail #117

Open
tobiHZDR opened this issue Jul 14, 2017 · 3 comments
Open

Convolution self test fail #117

tobiHZDR opened this issue Jul 14, 2017 · 3 comments
Assignees
Labels
Milestone

Comments

@tobiHZDR
Copy link

Compiling the master branch with CUDA 8.0 leads to a fail of the Convolution self test. On my system the bug arises in the testConvolution function. Outputting the Bsat from copypadmul2.cu gives wrong results in the case of the convolution self test. It seems like that the ToMSlice function produces the error. I also checked for the normal Demag calculation. There everything seems to be ok.

I used the standard problem for the test:
mumax3 -paranoid test/standardproblem4.mx3

Output from copypadmul2.cu
Bsat= 1.256637e-06 Ms= 2.338954e-313 Ms_mul= 1.000000e+00

And therefore,
convolution self-test tolerance: 0.37432534 FAIL

Can anyone reproduce this problem?

Best Regards,
Tobi

@godsic godsic added the bug label Jul 14, 2017
@godsic godsic self-assigned this Jul 14, 2017
@godsic
Copy link
Contributor

godsic commented Jul 14, 2017

@tobiHZDR I confirm the behavior with CUDA 8. Have you tried other CUDA versions?

@tobiHZDR
Copy link
Author

tobiHZDR commented Jul 14, 2017

@godsic I didn't try other CUDA versions. But today I can test CUDA 7.5

Edit: The problem also occurs using CUDA 7.5

@JonathanMaes
Copy link
Contributor

Bug still persists in CUDA 12.6.

@JonathanMaes JonathanMaes self-assigned this Oct 3, 2024
@JonathanMaes JonathanMaes added this to the 3.11 milestone Oct 10, 2024
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
Projects
None yet
Development

No branches or pull requests

3 participants