TIMEEV test fails in VASP 6.6.0

Questions regarding the compilation of VASP on various platforms: hardware, compilers and libraries, etc.


Moderators: Global Moderator, Moderator

Post Reply
Message
Author
john_low1
Newbie
Newbie
Posts: 22
Joined: Wed Nov 06, 2019 3:12 pm

TIMEEV test fails in VASP 6.6.0

#1 Post by john_low1 » Wed Mar 11, 2026 6:38 pm

To whom it may concern,

I have built VASP6.6.0 with gcc/13.2.0, openmpi/5.0.6, aocl/4.1.0, and hdf5/1.14.2 on an AMD EPYC 7713 64-Core Processor.

This run passes all tests in the testsuite except SiC_TIMEEV.
Note that the “errors” related to libxc and wannier90 are to be expected since neither libxc nor wannier90 was included in the build.

I noticed a known issue with TIMEEV and ISYM>O that would apply to this test. Is this the same issue or an error with my build?

I have attached my makefile.include, the log from the build, and the output from the testsuite.

Do you have any recommendations on how to resolve this issue?

John Low
Argonne National Laboratory.

You do not have the required permissions to view the files attached to this post.

alexey.tal
Global Moderator
Global Moderator
Posts: 510
Joined: Mon Sep 13, 2021 12:45 pm

Re: TIMEEV test fails in VASP 6.6.0

#2 Post by alexey.tal » Wed Mar 11, 2026 7:04 pm

Dear John,

I see that you have used 8 MPI ranks instead of 4. See the warning in the log file:

Code: Select all

Reference files have been generated with 4 MPI ranks.
Note that tests might fail if an other number of ranks is used!

Did you run the testsuite with 4 MPI ranks? Did the SiC_TIMEEV test pass?

Best regards,
Alexey


john_low1
Newbie
Newbie
Posts: 22
Joined: Wed Nov 06, 2019 3:12 pm

Re: TIMEEV test fails in VASP 6.6.0

#3 Post by john_low1 » Wed Mar 11, 2026 7:23 pm

Alexy,

I have run the SiC_TIMEEV test with 4 MPI ranks, and it passed!

Thank you!

John Low


alexey.tal
Global Moderator
Global Moderator
Posts: 510
Joined: Mon Sep 13, 2021 12:45 pm

Re: TIMEEV test fails in VASP 6.6.0

#4 Post by alexey.tal » Thu Mar 12, 2026 7:17 am

Glad to see that it solved the issue. I will investigate further why this test fails with 8 MPI ranks and see if we can make it more stable.


Post Reply