[theRock] bumping rocm-libraries blocked due to rocthrust OOM failure on Windows `gfx110X-all`

[HereThereBeDragons]

Last successful bumping was 5 days ago with commit 0588f07105ec85f6f36dd7f8e9bdd44603eece39

Recent try from yesterday ( ROCm/TheRock#1951 , rocm-library commit 487c14b0eeb61c60cb03ac925cba74c9894ac139 ) failed with

[rocThrust] [331/339] Building CXX object test/CMakeFiles/transform_scan.hip.dir/test_transform_scan.cpp.obj
[rocThrust] FAILED: test/CMakeFiles/transform_scan.hip.dir/test_transform_scan.cpp.obj 
[rocThrust] ccache B:\build\core\clr\dist\lib\llvm\bin\clang++.exe -DTHRUST_IGNORE_DEPRECATED_CPP_DIALECT -DUSE_PROF_API=1 -D__HIP_PLATFORM_AMD__=1 -IB:/build/math-libs/rocThrust/build/thrust/include -IC:/home/runner/_work/TheRock/TheRock/rocm-libraries/projects/rocthrust/thrust/.. -isystem B:/build/math-libs/rocThrust/build/test -isystem B:/build/math-libs/rocThrust/build/_deps/sqlite_local-src -isystem B:/build/math-libs/rocPRIM/dist/include -isystem B:/build/core/clr/dist/include -isystem B:/build/third-party/googletest/dist/include -DWIN32 -DWIN32_LEAN_AND_MEAN -D_CRT_SECURE_NO_WARNINGS -DNOMINMAX -fms-extensions -fms-compatibility -D_ENABLE_EXTENDED_ALIGNED_STORAGE  -Wno-documentation-unknown-command -Wno-documentation-pedantic -Wno-unused-command-line-argument -Wno-explicit-specialization-storage-class -Wno-ignored-attributes -Wno-unknown-attributes -Wno-duplicate-decl-specifier --hip-path=B:/build/core/clr/dist --hip-device-lib-path=B:/build/core/clr/dist/lib/llvm/amdgcn/bitcode -O3 -DNDEBUG -std=c++17 -D_DLL -D_MT -Xclang --dependent-lib=msvcrt -xhip -Wall -Wextra -x hip --offload-arch=gfx1100 --offload-arch=gfx1101 --offload-arch=gfx1102 --offload-arch=gfx1103 -MD -MT test/CMakeFiles/transform_scan.hip.dir/test_transform_scan.cpp.obj -MF test\CMakeFiles\transform_scan.hip.dir\test_transform_scan.cpp.obj.d -o test/CMakeFiles/transform_scan.hip.dir/test_transform_scan.cpp.obj -c C:/home/runner/_work/TheRock/TheRock/rocm-libraries/projects/rocthrust/test/test_transform_scan.cpp
[rocThrust] terminate called after throwing an instance of 'std::bad_alloc'

[rocThrust]   what():  std::bad_alloc

[rocThrust] [332/339] Building CXX object test/CMakeFiles/sort_permutation_iterator.hip.dir/test_sort_permutation_iterator.cpp.obj
[rocThrust] [333/339] Building CXX object test/CMakeFiles/sort.hip.dir/test_sort.cpp.obj
[rocThrust] [334/339] Building CXX object test/CMakeFiles/scan.hip.dir/test_scan.cpp.obj
[rocThrust] FAILED: test/CMakeFiles/scan.hip.dir/test_scan.cpp.obj 
[rocThrust] ccache B:\build\core\clr\dist\lib\llvm\bin\clang++.exe -DTHRUST_IGNORE_DEPRECATED_CPP_DIALECT -DUSE_PROF_API=1 -D__HIP_PLATFORM_AMD__=1 -IB:/build/math-libs/rocThrust/build/thrust/include -IC:/home/runner/_work/TheRock/TheRock/rocm-libraries/projects/rocthrust/thrust/.. -isystem B:/build/math-libs/rocThrust/build/test -isystem B:/build/math-libs/rocThrust/build/_deps/sqlite_local-src -isystem B:/build/math-libs/rocPRIM/dist/include -isystem B:/build/core/clr/dist/include -isystem B:/build/third-party/googletest/dist/include -DWIN32 -DWIN32_LEAN_AND_MEAN -D_CRT_SECURE_NO_WARNINGS -DNOMINMAX -fms-extensions -fms-compatibility -D_ENABLE_EXTENDED_ALIGNED_STORAGE  -Wno-documentation-unknown-command -Wno-documentation-pedantic -Wno-unused-command-line-argument -Wno-explicit-specialization-storage-class -Wno-ignored-attributes -Wno-unknown-attributes -Wno-duplicate-decl-specifier --hip-path=B:/build/core/clr/dist --hip-device-lib-path=B:/build/core/clr/dist/lib/llvm/amdgcn/bitcode -O3 -DNDEBUG -std=c++17 -D_DLL -D_MT -Xclang --dependent-lib=msvcrt -xhip -Wall -Wextra -x hip --offload-arch=gfx1100 --offload-arch=gfx1101 --offload-arch=gfx1102 --offload-arch=gfx1103 -MD -MT test/CMakeFiles/scan.hip.dir/test_scan.cpp.obj -MF test\CMakeFiles\scan.hip.dir\test_scan.cpp.obj.d -o test/CMakeFiles/scan.hip.dir/test_scan.cpp.obj -c C:/home/runner/_work/TheRock/TheRock/rocm-libraries/projects/rocthrust/test/test_scan.cpp
[rocThrust] terminate called after throwing an instance of 'std::bad_alloc'

[rocThrust]   what():  std::bad_alloc

a later run with a more recent commit had exactly the same failure.
https://github.com/ROCm/TheRock/actions/runs/18887990050/job/53932033094?pr=1955

Is this a known issue?

[HereThereBeDragons]

@jayhawk-commits @bstefanuk

[ScottTodd]

Commit range for rocprim: ROCm/rocPRIM@693b9bc...141eb19

[ScottTodd]

We have landed a workaround for this in TheRock by switching back from gfx110X-all (including gfx1103) to gfx110X-dgpu (not including it).

We suspect that enabling "offload compress" for all projects (and in particular rocprim and rocthrust) will give us more headroom for memory usage during the build.

[HereThereBeDragons]

closing due to workaround. underlying issue probably still exists with too many archs building at the same time running into resource exhaustion