Allow disabling the linking of libdevice in CUDACompilerParams

[gbaraldi]

Add option to disable the linking of libdevice into the IR. This is needed because https://github.com/EnzymeAD/Reactant.jl requires the intrinsics and not their linked code to raise LLVM IR into MLIR I imagine this can only be properly tested downstream

[wsmoses]

@gbaraldi one small thing : if the flag is set, can __nv_ functions as a legal intrinsic ? https://github.com/JuliaGPU/CUDA.jl/blob/c10e7cfb792b254ee17bf20afa9cc0bb5adca28d/src/compiler/compilation.jl#L25

[codecov[bot]]

Codecov Report

All modified and coverable lines are covered by tests :white_check_mark:

Project coverage is 59.95%. Comparing base (4d85f27) to head (d199899).

Additional details and impacted files

@@             Coverage Diff             @@
##           master    #2611       +/-   ##
===========================================
- Coverage   73.58%   59.95%   -13.64%     
===========================================
  Files         157      157               
  Lines       15242    15210       -32     
===========================================
- Hits        11216     9119     -2097     
- Misses       4026     6091     +2065     

:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.