LCAO buffer

Description

Change the buffer to be from single precision to be double precision.

This increases the memory footprint but makes that the results remain identical when changing the parallelization. The result of the test finite_systems_3d/01-carbon become identical for non-MPI and MPI builders when we do this.

Checklist

• I have checked that my code follows the Octopus coding standards
• I have added tests for all the new features added in this request.

changed milestone to %13.0

@micael.oliveira @martin.lueders What do you think about this? I know that this is maybe not ideal, but from the point of view of comparing results between bots, this is really nice as it eliminates one important source of discrepancy.

added 2 commits

• 2e13570d - Adding a save memory option for the LCAO.
• 0340fe5c - Updating some tests after the changes. Some testfiles now use the new option...

Compare with previous version

added 1 commit

• d91f4caf - Updating some tests after the changes. Some testfiles now use the new option...

Compare with previous version

added 1 commit

• 4590612a - Updating some tests after the changes. Some testfiles now use the new option...

Compare with previous version

@martin.lueders This merge request seems to be ok now. Interestingly, it seems that I found extra MPI versus non-MPI group of results in two tests, which are apparently not caused by the present problem. This will be my next step.

Once this is merged, I will also came back to the FFT grids (which originally revealed this problem).

added 25 commits

• 4590612a...dbc224e0 - 22 commits from branch main
• 4ae10702 - Change the buffer to be from single precision to be double precision.
• dc5cfacd - Adding a save memory option for the LCAO.
• cf927303 - Updating some tests after the changes. Some testfiles now use the new option...

Compare with previous version

@nicolastd This looks fine, as there is a variable to control the behavior.

approved this merge request

added 1 commit

• 063b68d2 - Updating some tests after the changes. Some testfiles now use the new option...

Compare with previous version

@martin.lueders Would you mind reapproving? I just updated two tolerances.

approved this merge request

added 1 commit

• 027c1fca - Updating some tests after the changes. Some testfiles now use the new option...

Compare with previous version

@martin.lueders This is now finally all green. Would you mind approving/merging this?

approved this merge request

enabled an automatic merge when the pipeline for 027c1fca succeeds

merged