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

libmusic + EmbeddedLapack + Cortex-M7 #2

Open
vaalberith opened this issue Aug 3, 2021 · 1 comment
Open

libmusic + EmbeddedLapack + Cortex-M7 #2

vaalberith opened this issue Aug 3, 2021 · 1 comment
Assignees
Labels
help wanted Extra attention is needed

Comments

@vaalberith
Copy link

vaalberith commented Aug 3, 2021

Greetings! I'm trying to make the mentioned combo work together, EmbeddedLapack contains f2c-converted blas and lapack libs. Tests with dgesvd_ (6, 7) keep failing, test 2,3 with dgesvd are ok. VT and S matrixes differs from reference ones in 6,7. Correlation matrix is ok everywhere. So it seems that dgesvd breaks with bigger matrixes OR data is incorrectly handled between correlation and dgesvd. Could you, please, tell, why should we shuffle data in such way (is it transpose or smth else)? Sorry for stupid questions and thanks in advance!
for (i = 0; i < M; ++i) { for (j = 0; j < N; ++j) { A[j * M + i] = Y[i * yn + j]; } }

@piotrgregor piotrgregor self-assigned this Aug 3, 2021
@piotrgregor piotrgregor added the help wanted Extra attention is needed label Aug 3, 2021
@piotrgregor
Copy link
Contributor

Thank you very much for your comments @vaalberith Both lmtest2.c and lmtest6.c use dgesvd_ to compute job parameters and job result. And yes:
for (i = 0; i < M; ++i) { for (j = 0; j < N; ++j) { A[j * M + i] = Y[i * yn + j]; } }
is to transpose Y[M, N] into A[N, M] cause this is what dgesvd_ is expecting.

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
help wanted Extra attention is needed
Projects
None yet
Development

No branches or pull requests

2 participants