[flang][rt] Add noinline attributes for CUDA compile path for successful compilation

[modiking]

NVCC does more aggressive inlining than Clang/GCC causing the exported functions in extrema.cpp and findloc.cpp to become extremely large from function specializations leading to compilation timeouts. Marking the 2 functions in this change as noinline for NVCC alleviates this problem as it removes the worst of the cross-matrix argument specializations.

Also remove the workaround in #156542 that opted out findloc.cpp from the CUDA flang-rt build

Testing:
ninja flang-rt builds in ~30 minutes, these 2 files build in ~3 minutes

[clementval]

Thanks for the workaround @modiking. I added @klausler who is the main runtime developer as reviewer. Give him some time to chime in if he has anything to say.

[github-actions]

✅ With the latest revision this PR passed the C/C++ code formatter.

[vzakhari]

Thanks for the changes! Please use RT_DEVICE_NOINLINE macro instead. We have it defined in flang/include/flang/Common/api-attrs.h.

[clementval]

LGTM. Thanks

[klausler]

Please get Slava's approval before merging.

[vzakhari]

Thank you!

[modiking]

Appreciate the quick reviews!