[Chapel Merge] Allow CHPL_LOCALE_MODEL=gpu and CHPL_COMM=gasnet

Branch: refs/heads/main
Revision: e2e2bc5
Author: stonea
Link: Unavailable
Log Message:

Merge pull request #19857 from stonea/use_gasnet_no_refactor

Allow CHPL_LOCALE_MODEL=gpu and CHPL_COMM=gasnet

This PR isn't intended to enable communication from/to the GPU across nodes but (the hope atleast) is that you'll be able to do the normal node-to-node communication you're used to.

[Reviewed by @ronawho and @e-kayrakli]

Modified Files:
A test/gpu/native/distArray/SKIPIF

M compiler/codegen/codegen.cpp
M compiler/optimizations/gpuTransforms.cpp
M modules/internal/LocaleModelHelpSetup.chpl
M test/gpu/native/environment/gasnet.chpl
M test/gpu/native/environment/gasnet.good
M third-party/qthread/Makefile
M util/chplenv/chpl_gpu.py

Compare: https://github.com/chapel-lang/chapel/compare/386292edf966...e2e2bc5c376f