[REVIEW] Parallelize Treelite to FIL conversion over trees

[wphicks]

Convert Treelite trees to FIL in parallel
Improve speed of initial prediction for a model with 300 trees of approximately 4M nodes by an average of 5.9% on a system with 12 Intel Xeon Gold 6128 CPUs available

[wphicks]

Note that @canonizer previously proposed replacing std::vector for storing FIL nodes with a raw malloc call to avoid the initialization overhead. Profiling demonstrates that this provides only a 0.5% speedup relative to an initial resize call on the vector, which makes sense given the constructors for the FIL nodes. Furthermore, looking at the generated assembly for each approach on Godbolt, I'm fairly well convinced that the vast majority of that tiny slowdown when using std::vector is genuinely useful error handling. Given that, I think it's worth it to accept the slowdown in favor of better error-handling and maintainability. If we decide we do want that extra 0.5%, we can use make_unique to generate the array and avoid the raw malloc and free, as demonstrated in an earlier commit on this branch.

[codecov-io]

Codecov Report

Merging #3396 (92a1548) into branch-0.18 (550121b) will increase coverage by 0.14%.
The diff coverage is 85.77%.

@@               Coverage Diff               @@
##           branch-0.18    #3396      +/-   ##
===============================================
+ Coverage        71.48%   71.63%   +0.14%     
===============================================
  Files              207      210       +3     
  Lines            16748    16924     +176     
===============================================
+ Hits             11973    12123     +150     
- Misses            4775     4801      +26     

Impacted Files	Coverage Δ
python/cuml/decomposition/incremental_pca.py	94.70% <ø> (ø)
python/cuml/dask/ensemble/base.py	19.69% <30.43%> (+0.36%)	⬆️
python/cuml/ensemble/randomforestregressor.pyx	70.83% <44.44%> (ø)
...ython/cuml/dask/ensemble/randomforestclassifier.py	30.00% <50.00%> (+0.51%)	⬆️
python/cuml/dask/ensemble/randomforestregressor.py	35.08% <50.00%> (+0.54%)	⬆️
python/cuml/fil/fil.pyx	91.87% <60.00%> (-1.88%)	⬇️
python/cuml/ensemble/randomforestclassifier.pyx	73.72% <66.66%> (ø)
python/cuml/multiclass/multiclass.py	84.21% <84.21%> (ø)
python/cuml/model_selection/_split.py	90.35% <90.35%> (ø)
python/cuml/svm/svm_base.pyx	94.27% <91.30%> (-0.63%)	⬇️
... and 17 more

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update c48f35a...92a1548. Read the comment docs.

[canonizer]

Just small technical comments, otherwise looks good.

[canonizer]

Approved again.

[JohnZed]

@gpucibot merge