Modify the interface of the tpgqrt routines, and fix some numerical bug within the routine

@all if anyone can check that I didn't miss anything.

added Bug label

changed milestone to %Chameleon 1.0.0

assigned to @faverge

added 1 commit

• 396b456f - Remove debug check

Compare with previous version

Thanks for the notification.

"This build is stuck, because the project doesn't have any runners online assigned to it. Go to Runners page "

You changed something in your configuration, I think the previous merge request build correctly ?

Thanks. It works better now :).

@ltaief Please check if everything looks fine to you in this pull request, and if so, click on the "thumb up" at the top of the discussion. Thanks

@faverge I mostly looked at the testing code for now (testing_dgeqrt_qdwh), you do rebuild Q from Q1 and Q2 but I cannot see the rebuild of A from A1 and A2.

Line 104:
/* Initialize A1, A2, and A */ LAPACKE_zlarnv_work(IONE, ISEED, MxM, A1); LAPACKE_zlaset_work( LAPACK_COL_MAJOR, 'A', M, M, 0., 1., A2, M );

LAPACKE_zlacpy_work( LAPACK_COL_MAJOR, 'A', M, M, A1, M, A,     LDA );
LAPACKE_zlacpy_work( LAPACK_COL_MAJOR, 'A', M, M, A2, M, A + M, LDA );

I initialize A1 random, A2 to Id as in QDWH (I should change that for a more generic code maybe), and backup A1/A2 in A. Is that what you are talking about?

Yes it is. I have probably looked then at a previous version. Looks good!

added 8 commits

• 396b456f...14486691 - 3 commits from branch solverstack:master
• 6826cefc - Fix parameters in tpqrt/tpmqrt starpu codelets
• ce8be021 - Change tpgqrt interface because ungqr already includes laset, so it can not be u…
• 8e9a7503 - Fix check on triangular part of pentagonal matrices
• 0825038b - Add a testing for the QR factorization of two square matrices on top of each oth…
• 65aadc18 - Remove debug check

Compare with previous version

enabled an automatic merge when the pipeline for 65aadc18 succeeds

merged

mentioned in commit 4b3a2f82