================
@@ -322,22 +322,26 @@ foreach( t ${LIBCLC_TARGETS_TO_BUILD} )
if ( ARCH STREQUAL spirv OR ARCH STREQUAL spirv64 )
set( build_flags -O0 -finline-hint-functions -DCLC_SPIRV )
- set( opt_flags )
+ set( clc_opt_flags )
+ # Inline CLC functions into OpenCL builtins
+ set( opencl_opt_flags "--passes=always-inline" )
----------------
arsenm wrote:
The correct IR implementation of FMA is universally call the llvm.fma
intrinsic. There's no need for additional configuration in libclc. The choice
to emit a call or whatnot is then a codegen decision, and is subject to
nobuiltin-like configuration
https://github.com/llvm/llvm-project/pull/115699
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
