Add basic support to CUDA

[mborland]

No description provided.

[mborland]

Build errors should be fixed by: boostorg/math#1168

[codecov]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 99.1%. Comparing base (4bf4e48) to head (f3bbbb4).

Additional details and impacted files

@@            Coverage Diff            @@
##           develop    #717     +/-   ##
=========================================
- Coverage     99.1%   99.1%   -0.0%     
=========================================
  Files          230     230             
  Lines        16431   16432      +1     
  Branches      1779    1781      +2     
=========================================
- Hits         16279   16278      -1     
- Misses         152     154      +2     

Files	Coverage Δ
include/boost/decimal/decimal32_fast.hpp	96.8% <100.0%> (ø)
include/boost/decimal/detail/apply_sign.hpp	100.0% <100.0%> (ø)
include/boost/decimal/detail/fenv_rounding.hpp	100.0% <100.0%> (ø)
...lude/boost/decimal/detail/integer_search_trees.hpp	100.0% <100.0%> (ø)
include/boost/decimal/detail/normalize.hpp	100.0% <100.0%> (ø)

... and 1 file with indirect coverage changes

---

Continue to review full report in Codecov by Sentry.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 4bf4e48...f3bbbb4. Read the comment docs.

[mborland]

Back to green now that upstream math is fixed

[mborland]

This gets add, sub, mul, and div working for decimal32_fast on CUDA. Many more PRs to follow I'm sure

[mborland]

Upstream fixes to math have been merged. Will double back if there's any remaining failures but I expect there will not be.