Improve mpi buffer allocation in post_process

[hyeoksu-lee]

Description

In the subroutine s_initialize_mpi_proxy_module in post_process, q_cons_buffer_in and q_cons_buffer_out are allocated with the size buff_size*sys_size*(m + 2*buff_size + 1)*(n + 2*buff_size + 1)*(p + 2*buff_size + 1)/(min(m, n, p)+ 2*buff_size + 1). However, as the Fortran computes the operations sequentially, it could be too large before dividing by (min(m, n, p)+ 2*buff_size + 1).

For example, I noticed in a 3D simulation (1023 x 1023 x 511) with method of classes bubbles (nb = 11) that buff_size*sys_size*(m + 2*buff_size + 1)*(n + 2*buff_size + 1)*(p + 2*buff_size + 1) is 6.4242e+10, which is greater than the max value of integer 2^31 - 1 =2.1475e+09. Therefore, it becomes a wrong negative integer value, -2528337, and thus it throws an error.

Therefore, I made a quick fix to this by changing the order of operations.

Type of change

Please delete options that are not relevant.

• Bug fix (non-breaking change which fixes an issue)

Scope

• This PR comprises a set of related changes with a common goal

If you cannot check the above box, please split your PR into multiple PRs that each have a common goal.

How Has This Been Tested?

• Large 3D simulation (1023 x 1023 x 511, sys_size = 29)

• What computers and compilers did you use to test this: Carpenter

Checklist

• I ran ./mfc.sh format before committing my code
• This PR does not introduce any repeated code (it follows the DRY principle)
• I cannot think of a way to condense this code and reduce any introduced additional line count

[sbryngelson]

Good find. There's actually a "better" way of doing this by making them floating point additions/multiplies then turning the result into the nearest integer. @wilfonba can you incorporate this fix into your MPI PR?

[codecov]

Codecov Report

Attention: Patch coverage is 0% with 2 lines in your changes missing coverage. Please review.

Project coverage is 43.09%. Comparing base (40164cf) to head (518a645).

Files with missing lines	Patch %	Lines
src/post_process/m_mpi_proxy.fpp	0.00%	2 Missing ⚠️

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #837      +/-   ##
==========================================
- Coverage   43.15%   43.09%   -0.07%     
==========================================
  Files          68       68              
  Lines       20262    20262              
  Branches     2424     2427       +3     
==========================================
- Hits         8745     8731      -14     
- Misses      10054    10071      +17     
+ Partials     1463     1460       -3     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.

[wilfonba]

A fix for this problem is already in the MPI refactor PR (and this code in post-process is removed in favor of a common module) .https://github.com/wilfonba/MFC-Wilfong/blob/6448b3701fa44700b99955107ddb27f961ce5749/src/common/m_mpi_common.fpp#L54-L81
I'll hopefully have time to finish up the necessary manual checks for that PR this weekend once I get back to Atlanta.

[hyeoksu-lee]

@sbryngelson @wilfonba Sounds great. I will close this PR then. Thanks!