ENH: Annotate mris_expand with thread usage

[effigies]

This one's a little weird, due to the multiple inheritance and also the inefficiency of multithreading in mris_expand.

I noticed midthickness taking a long time while running tests, and I was pretty sure I'd seen that it could be parallelized.

From watching htop and running mris_expand with a set of OMP_NUM_THREADS values, I see that generally you get about 65-75% core utilization, generally a bit under. We therefore overallocate threads by about 50% to make the core usage match the desired number of threads, occasionally peaking over but mostly sitting a bit under. This works pretty well for 2-12 desired threads (3-18 OMP_NUM_THREADS), but seems almost completely limited after that, with 24 threads utilizing maybe 13 cores at best. Therefore, I set this to max out at 12 for scheduling purposes.

The memory usage of this tool is low and does not seem to increase noticeably with number of threads.

If n_procs == 1, str(1 * 3 // 2) == '1', so this does not interfere with people attempting to run things single-threaded.

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (b5e8663) 53.47% compared to head (43eed0e) 58.48%.

❗ Current head 43eed0e differs from pull request most recent head 899f2f4. Consider uploading reports for the commit 899f2f4 to get more accurate results

Additional details and impacted files

@@            Coverage Diff             @@
##             next     #386      +/-   ##
==========================================
+ Coverage   53.47%   58.48%   +5.00%     
==========================================
  Files          24       27       +3     
  Lines        1840     1910      +70     
  Branches      239      240       +1     
==========================================
+ Hits          984     1117     +133     
+ Misses        787      736      -51     
+ Partials       69       57      -12     

Flag	Coverage Δ
ds005	∅ <ø> (?)
ds054	45.99% <71.42%> (+0.06%)	⬆️
pytest	37.12% <100.00%> (?)

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

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

[effigies]

If there are no objections, I'll merge this before a release. It should reduce a 20min job to 2.5-4 minutes.