Cuda request gpu device expr

[tetron]

• Revise CUDARequirement so to support expressions to choose how many GPUs we want.
• adjust cudaComputeCapability to take an array of explicit versions to use

[codecov]

Codecov Report

Merging #1629 (3fd94f3) into main (4c65bdf) will increase coverage by 0.11%.
The diff coverage is 73.91%.

@@            Coverage Diff             @@
##             main    #1629      +/-   ##
==========================================
+ Coverage   66.73%   66.85%   +0.11%     
==========================================
  Files          93       93              
  Lines       16566    16580      +14     
  Branches     4396     4404       +8     
==========================================
+ Hits        11056    11085      +29     
+ Misses       4374     4357      -17     
- Partials     1136     1138       +2     

Impacted Files	Coverage Δ
cwltool/docker.py	71.54% <0.00%> (+0.17%)	⬆️
cwltool/executors.py	82.58% <0.00%> (-0.75%)	⬇️
cwltool/singularity.py	58.21% <0.00%> (-0.20%)	⬇️
cwltool/cuda.py	75.00% <66.66%> (+39.70%)	⬆️
cwltool/job.py	79.88% <100.00%> (+1.22%)	⬆️
cwltool/process.py	88.94% <100.00%> (+0.15%)	⬆️
cwltool/cwltool/executors.py	64.73% <0.00%> (-0.59%)	⬇️
cwltool/cwltool/process.py	70.96% <0.00%> (-0.53%)	⬇️
cwltool/cwltool/singularity.py	52.11% <0.00%> (-0.23%)	⬇️
cwltool/cwltool/docker.py	49.37% <0.00%> (-0.01%)	⬇️
... and 4 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 4c65bdf...3fd94f3. Read the comment docs.

[lgtm-com]

This pull request introduces 2 alerts when merging a30b00b into 4c65bdf - view on LGTM.com

new alerts:

• 2 for Unused import

[lgtm-com]

This pull request introduces 2 alerts when merging 090b7e6 into 4c65bdf - view on LGTM.com

new alerts:

• 2 for Unused import

[lgtm-com]

This pull request introduces 2 alerts when merging 3fd94f3 into 4c65bdf - view on LGTM.com

new alerts:

• 2 for Unused import