Fix QChem not using max_core

[t-young31]

Resolves #273

Hopefully this is correct!

---

Checklist

• [x] The changes include an associated explanation of how/why
• [ ] Test pass
• [x] Documentation has been updated
• [x] Changelog has been updated

[codecov[bot]]

Codecov Report

Merging #275 (cf8a7cf) into v1.4.0 (6210a14) will decrease coverage by 0.05%. The diff coverage is 100.00%.

@@            Coverage Diff             @@
##           v1.4.0     #275      +/-   ##
==========================================
- Coverage   97.24%   97.19%   -0.05%     
==========================================
  Files         197      197              
  Lines       21013    21026      +13     
==========================================
+ Hits        20434    20437       +3     
- Misses        579      589      +10     

Flag	Coverage Δ
unittests	97.19% <100.00%> (-0.05%)	:arrow_down:

Flags with carried forward coverage won't be shown. Click here to find out more.

Impacted Files	Coverage Δ
autode/wrappers/QChem.py	97.49% <100.00%> (+0.01%)	:arrow_up:
tests/test_wrappers/test_qchem.py	100.00% <100.00%> (ø)

... and 1 file with indirect coverage changes

[t-young31]

Thanks @shoubhikraj – I've had a look at the manual now too and I'm not really any the wiser. Seems like the default CC_MEMORY is 50% MEM_TOTAL, but is that in addition to MEM_TOTAL?! @asterlingchem any ideas?

[asterlingchem]

From my interpretation of https://manual.q-chem.com/latest/Ch6.S16.SS2.html and https://manual.q-chem.com/latest/Ch6.S16.SS4.html, when using the default CCMAN2 library, CC_MEMORY is set to 50%* of MEM_TOTAL, so you shouldn't need to specify CC_MEMORY explicitly in most cases (assuming the default has been sensibly chosen). It's a similar story for MEM_STATIC – this value is set automatically as 12% of MEM_TOTAL. So as long as MEM_TOTAL is set, and is appropriate for the resources available on the node, then CC_MEMORY and MEM_STATIC shouldn't need to be specified.

*This is a bit weird because the recommendation from 6.16.2 of the manual is to use 75-80% of MEM_TOTAL, but this appears to be for exclusive node use – so maybe the 50% is a safer threshold?

[t-young31]

awesome – thanks @asterlingchem