parallelize some csg operations

[pca006132]

Just some trivial parallelization, results in ~5% improvement for meshes with many unions. Just thought of this and this makes me unable to fall asleep...
BatchBoolean is not optimized yet, not sure if I should do parallelization as it will mess up with the mesh size and may end up being slower.

[pca006132]

Btw @elalish part of the compose function has to be executed sequentially and will cause segfault when parallelized (the for loop that I left), not sure if that is a bug.

[codecov]

Codecov Report

Patch coverage: 94.18% and project coverage change: +0.48 🎉

Comparison is base (d713986) 85.48% compared to head (a77984a) 85.96%.

Additional details and impacted files

@@            Coverage Diff             @@
##           master     #381      +/-   ##
==========================================
+ Coverage   85.48%   85.96%   +0.48%     
==========================================
  Files          36       36              
  Lines        4415     4468      +53     
==========================================
+ Hits         3774     3841      +67     
+ Misses        641      627      -14     

Impacted Files	Coverage Δ
src/cross_section/src/cross_section.cpp	71.03% <86.20%> (+9.70%)	⬆️
src/manifold/src/csg_tree.cpp	92.33% <98.11%> (+0.25%)	⬆️
src/manifold/src/sort.cpp	90.37% <100.00%> (ø)
src/utilities/include/par.h	94.28% <100.00%> (+0.53%)	⬆️

... and 2 files with indirect coverage changes

Help us with your feedback. Take ten seconds to tell us how you rate us. Have a feature suggestion? Share it here.

☔ View full report in Codecov by Sentry.
📢 Do you have feedback about the report comment? Let us know in this issue.

[elalish]

Thanks!