[STF] Option to disable kernel generation in CUDASTF (NVIDIA#2723)

* Add an option to control whether we can generate CUDA kernels with parallel_For and launch
fbusato · Nov 9, 2024 · 39ba5f5 · 39ba5f5
1 parent 52ea4e6
commit 39ba5f5
Showing 1 changed file with 1 addition and 0 deletions.
diff --git a/cudax/CMakeLists.txt b/cudax/CMakeLists.txt
@@ -21,6 +21,7 @@ option(cudax_ENABLE_HEADER_TESTING "Test that CUDA Experimental's public headers
 option(cudax_ENABLE_TESTING "Build CUDA Experimental's tests." ON)
 option(cudax_ENABLE_EXAMPLES "Build CUDA Experimental's examples." ON)
 option(cudax_ENABLE_CUDASTF "Enable CUDASTF subproject" ON)
+option(cudax_ENABLE_CUDASTF_CODE_GENERATION "Enable code generation using STF's parallel_for or launch with CUDA compiler." ON)
 option(cudax_ENABLE_CUDASTF_BOUNDSCHECK "Enable bounds checks for STF targets. Requires debug build." OFF)
 option(cudax_ENABLE_CUDASTF_DEBUG "Enable additional debugging for STF targets. Requires debug build." OFF)
 option(cudax_ENABLE_CUDASTF_MATHLIBS "Enable STF tests/examples that use cublas/cusolver." OFF)