CUDA: Generic all and all-major support

Commit 14d8a276 (CUDA: Support nvcc 11.5 new -arch=all|all-major flags,
2021-08-17) added all and all-major options to CUDA_ARCHITECTURES. These are
fairly generic and likely to see real-world use by distributors. Thus it's
desirable to support these also for Clang and older NVCC versions.

The supported architectures are dependent on the toolkit version. We determine
the toolkit version prior to compiler detection. For NVCC we get the version
from the vendor identification output, but for Clang we need to invoke NVCC
separately.

The architecture information is mostly based on the Wikipedia list with the
earliest supported version being CUDA 7.0. This could be documented and
expanded in the future to allow projects to query CUDA toolkit version and
architecture information.
For Clang we additionally constrain based on its support.

Additionally the architecture mismatch detection logic is fixed, improved and
updated for generic support:
* Commit 01428c55 (CUDA: Fail fast if CMAKE_CUDA_ARCHITECTURES doesn't work
  during detection, 2020-08-29) enabled CMAKE_CUDA_COMPILER_ID_REQUIRE_SUCCESS
  if CMAKE_CUDA_ARCHITECTURES is specified. This results in
  CMakeDetermineCompilerID.cmake printing the compiler error and our code for
  presenting the mismatch in a user-friendly way being useless. The custom
  logic seems preferable so go back to not enabling it.
* Commit 14d8a276 (CUDA: Support nvcc 11.5 new -arch=all|all-major flags,
  2021-08-17) tried to support CMP0054 but forgot to add x to the interpolated
  result. Thus the conditions would always evaluate to false. This is fixed as
  a byproduct of removing NVIDIA specific checks, improving the error message
  and replacing architectures_mode with a simpler architectures_explicit.

Visual Studio support omits testing the flags during detection due to
complexities in determining the toolkit version when using it.
A long-term proper implementation would be #23161.

Implements #22860.

1 files changed, 46 insertions, 34 deletions

diff --git a/Tests/CudaOnly/All/CMakeLists.txt b/Tests/CudaOnly/All/CMakeLists.txt
index fe29bb0..ba32e9a 100644
--- a/Tests/CudaOnly/All/CMakeLists.txt
+++ b/Tests/CudaOnly/All/CMakeLists.txt
@@ -2,43 +2,55 @@ cmake_minimum_required(VERSION 3.20)
 project(CudaOnlyAll CUDA)
 
 if(CMAKE_CUDA_COMPILER_ID STREQUAL "NVIDIA" AND
-   CMAKE_CUDA_COMPILER_VERSION VERSION_GREATER_EQUAL 11.5.0)
-
+   CMAKE_CUDA_COMPILER_VERSION VERSION_GREATER_EQUAL 8.0)
   set(compile_options -Wno-deprecated-gpu-targets)
-  function(verify_output flag output_var)
-    string(REGEX MATCHALL "-arch compute_([0-9]+)" target_archs "${${output_var}}")
-    list(LENGTH target_archs count)
-    if(count LESS 2)
-      message(FATAL_ERROR "${flag} failed to map to multiple architectures")
-    endif()
-  endfunction()
 endif()
 
-if(COMMAND verify_output)
-  set(try_compile_flags -v ${compile_options})
-
-  set(CMAKE_CUDA_ARCHITECTURES all)
-  try_compile(all_archs_compiles
-    ${CMAKE_CURRENT_BINARY_DIR}/try_compile/all_archs_compiles
-    ${CMAKE_CURRENT_SOURCE_DIR}/main.cu
-    COMPILE_DEFINITIONS ${try_compile_flags}
-    OUTPUT_VARIABLE output
-    )
-  verify_output(all output)
-
-  set(CMAKE_CUDA_ARCHITECTURES all-major)
-  try_compile(all_major_archs_compiles
-    ${CMAKE_CURRENT_BINARY_DIR}/try_compile/all_major_archs_compiles
-    ${CMAKE_CURRENT_SOURCE_DIR}/main.cu
-    COMPILE_DEFINITIONS ${try_compile_flags}
-    OUTPUT_VARIABLE output
-    )
-  verify_output(all-major output)
-
-  if(all_archs_compiles AND all_major_archs_compiles)
-    add_executable(CudaOnlyAll main.cu)
-    target_compile_options(CudaOnlyAll PRIVATE ${compile_options})
+function(verify_output flag)
+  string(REPLACE "-" "_" architectures "${flag}")
+  string(TOUPPER "${architectures}" architectures)
+  set(architectures "${CMAKE_CUDA_ARCHITECTURES_${architectures}}")
+
+  if(CMAKE_CUDA_COMPILER_ID STREQUAL "Clang")
+    set(match_regex "-target-cpu sm_([0-9]+)")
+  elseif(CMAKE_CUDA_COMPILER_ID STREQUAL "NVIDIA")
+    set(match_regex "-arch compute_([0-9]+)")
+  endif()
+
+  string(REGEX MATCHALL "${match_regex}" target_cpus "${output}")
+
+  foreach(cpu ${target_cpus})
+    string(REGEX MATCH "${match_regex}" dont_care "${cpu}")
+    list(APPEND command_archs "${CMAKE_MATCH_1}")
+  endforeach()
+
+  list(SORT command_archs)
+  if(NOT "${command_archs}" STREQUAL "${architectures}")
+    message(FATAL_ERROR "Architectures used for \"${flag}\" don't match the reference (\"${command_archs}\" != \"${architectures}\").")
   endif()
-else()
+endfunction()
+
+set(try_compile_flags -v ${compile_options})
+
+set(CMAKE_CUDA_ARCHITECTURES all)
+try_compile(all_archs_compiles
+  ${CMAKE_CURRENT_BINARY_DIR}/try_compile/all_archs_compiles
+  ${CMAKE_CURRENT_SOURCE_DIR}/main.cu
+  COMPILE_DEFINITIONS ${try_compile_flags}
+  OUTPUT_VARIABLE output
+  )
+verify_output(all)
+
+set(CMAKE_CUDA_ARCHITECTURES all-major)
+try_compile(all_major_archs_compiles
+  ${CMAKE_CURRENT_BINARY_DIR}/try_compile/all_major_archs_compiles
+  ${CMAKE_CURRENT_SOURCE_DIR}/main.cu
+  COMPILE_DEFINITIONS ${try_compile_flags}
+  OUTPUT_VARIABLE output
+  )
+verify_output(all-major)
+
+if(all_archs_compiles AND all_major_archs_compiles)
   add_executable(CudaOnlyAll main.cu)
+  target_compile_options(CudaOnlyAll PRIVATE ${compile_options})
 endif()