General benchmarking framework

[b-fg]

Automated benchmarks generation to improve in reproducibility, and profiling of new code.
When completed, it should be able to automatically run and compare benchmarks across multiple:

• Julia versions
• Backends
• Cases (including size, simulation time/max. steps, data type)

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Comparison is base (ff83ce9) 89.47% compared to head (ba88cc9) 91.07%.
Report is 11 commits behind head on master.

Additional details and impacted files

@@            Coverage Diff             @@
##           master      #86      +/-   ##
==========================================
+ Coverage   89.47%   91.07%   +1.60%     
==========================================
  Files           9        9              
  Lines         399      538     +139     
==========================================
+ Hits          357      490     +133     
- Misses         42       48       +6     

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

[weymouth]

Looks cool. A couple things.

• We probably need to add documentation on how to use this.
• This adds dependencies which aren't needed for just running the code. Can we make these optional by including them as part of a new target? (I've been thinking we should try to reduce/organize our dependencies better.)

[b-fg]

Yes, I was planning to look into better handling the extra dependencies too. And documentation will also come in due course :)

[b-fg]

Pretty much done :) Feel free to give it a try!

• Documentation can be found within the benchmark.sh file.
• Fixed project dependencies by moving them to a subproject environment.

Output example on the TGV case:

[b-fg]

I think this is finished @weymouth. We can add more benchmark examples in the future, additionally to the tgv.jl.
After merging this PR, and figuring out optional dependencies we could release the minor update.