[Libomptarget] Replace Nvidia arch lookup with 'nvptx-arch'

[jhuber6]

Summary

This method to look up the CUDA architecture is deprecated in newer
versions of CMake. We also have our own way to query this information
that we control now via the nvptx-arch program, which should always be
present in LLVM builds with clang going forward. This is currently only
used for testing so I think we should be okay with the dependency.

[jhuber6]

Fix conditional in case the program isn't found.

[tianshilei1992]

openmp/libomptarget/cmake/Modules/LibomptargetGetDependencies.cmake

107

We also have to update this to FindCUDAToolkit.

openmp/libomptarget/test/lit.cfg

88

why this?

[jhuber6]

openmp/libomptarget/cmake/Modules/LibomptargetGetDependencies.cmake

107

We can't do that until the LLVM CMake version is bumped up in the 17 branch.

[jhuber6]

openmp/libomptarget/test/lit.cfg

88

The indentation was an accident. The slice is because the old version returned just 70 while the new one would output sm_70.

[jhuber6]

Fix incorrect indentation.

[tianshilei1992]

LG

openmp/libomptarget/cmake/Modules/LibomptargetGetDependencies.cmake

mark_as_advanced(
   LIBOMPTARGET_DEP_LIBFFI_LIBRARIES)
 
 ################################################################################
 # Looking for CUDA...
 ################################################################################
 if (CUDA_TOOLKIT_ROOT_DIR)
   set(LIBOMPTARGET_CUDA_TOOLKIT_ROOT_DIR_PRESET TRUE)
 endif()
 find_package(CUDA QUIET)
 
-# Try to get the highest Nvidia GPU architecture the system supports
-if (CUDA_FOUND)
-  cuda_select_nvcc_arch_flags(CUDA_ARCH_FLAGS)
-  string(REGEX MATCH "sm_([0-9]+)" CUDA_ARCH_MATCH_OUTPUT ${CUDA_ARCH_FLAGS})
-  if (NOT DEFINED CUDA_ARCH_MATCH_OUTPUT OR "${CMAKE_MATCH_1}" LESS 35)
-    libomptarget_warning_say("Setting Nvidia GPU architecture support for OpenMP target runtime library to sm_35 by default")
-    set(LIBOMPTARGET_DEP_CUDA_ARCH "35")
-  else()
-    set(LIBOMPTARGET_DEP_CUDA_ARCH "${CMAKE_MATCH_1}")
+# Identify any locally installed GPUs to use for testing.
+set(LIBOMPTARGET_DEP_CUDA_ARCH "sm_35")
+
+find_program(LIBOMPTARGET_NVPTX_ARCH NAMES nvptx-arch PATHS ${LLVM_BINARY_DIR}/bin)
+if(LIBOMPTARGET_NVPTX_ARCH)
+  execute_process(COMMAND ${LIBOMPTARGET_NVPTX_ARCH}
+                  OUTPUT_VARIABLE LIBOMPTARGET_NVPTX_ARCH_OUTPUT
+                  OUTPUT_STRIP_TRAILING_WHITESPACE)
+  string(FIND "${LIBOMPTARGET_NVPTX_ARCH_OUTPUT}" "\n" first_arch_string)
+  string(SUBSTRING "${LIBOMPTARGET_NVPTX_ARCH_OUTPUT}" 0 ${first_arch_string}
+         arch_string)
+  if(arch_string)
+    set(LIBOMPTARGET_DEP_CUDA_ARCH "${arch_string}")
   endif()
 endif()
 
 set(LIBOMPTARGET_DEP_CUDA_FOUND ${CUDA_FOUND})
 set(LIBOMPTARGET_DEP_CUDA_INCLUDE_DIRS ${CUDA_INCLUDE_DIRS})
 
 mark_as_advanced(
   LIBOMPTARGET_DEP_CUDA_FOUND

openmp/libomptarget/test/lit.cfg

 config.available_features.add(config.libomptarget_current_target)
 
 # Determine whether the test system supports unified memory.
 # For CUDA, this is the case with compute capability 70 (Volta) or higher.
 # For all other targets, we currently assume it is.
 supports_unified_shared_memory = True
 if config.libomptarget_current_target.startswith('nvptx'):
   try:
-    cuda_arch = int(config.cuda_test_arch)
+    cuda_arch = int(config.cuda_test_arch[:3])
     if cuda_arch < 70:
       supports_unified_shared_memory = False
   except ValueError:
     # If the architecture is invalid, assume it is supported.
     supports_unified_shared_memory = True
 if config.libomptarget_current_target.startswith('amdgcn'):
     supports_unified_shared_memory = False
 if supports_unified_shared_memory: