WIP: Refactor serac to support RAJA

[johnbowen42]

This PR enables CUDA execution via RAJA, as well as compilation of GPU kernels with NVCC. Several important restrictions imposed by nvcc that this PR works around are
(1) Multiple variadic template parameters in the enclosing function of a generic lambda
(2) Using [=] to capture local variables in a lambda declaration doesn't work with if the enclosing function has a variadic template parameter
(3) Fully generic lambdas (where both arguments and return type are deduced with auto)

Some template programming workarounds were required to expand the multiple template parameter restriction of nvcc, but this is currently passing some very basic unit tests ported from CPU.

[codecov-commenter]

Codecov Report

Merging #987 (26cf13c) into develop (b10a5fb) will decrease coverage by 0.17%.
Report is 43 commits behind head on develop.
The diff coverage is 100.00%.

@@             Coverage Diff             @@
##           develop     #987      +/-   ##
===========================================
- Coverage    89.35%   89.19%   -0.17%     
===========================================
  Files          138      138              
  Lines        10722    10830     +108     
===========================================
+ Hits          9581     9660      +79     
- Misses        1141     1170      +29     

Files Changed	Coverage Δ
...ac/numerics/functional/domain_integral_kernels.hpp	100.00% <100.00%> (ø)
src/serac/physics/heat_transfer.hpp	86.42% <100.00%> (+0.12%)	⬆️

... and 3 files with indirect coverage changes

📣 We’re building smart automated test selection to slash your CI/CD build times. Learn more

📢 Have feedback on the report? Share it here.