================
@@ -337,9 +337,12 @@ void AMDGPUTargetInfo::getTargetDefines(const LangOptions
&Opts,
if (hasFastFMA())
Builder.defineMacro("FP_FAST_FMA");
- Builder.defineMacro("__AMDGCN_WAVEFRONT_SIZE__", Twine(WavefrontSize));
- // ToDo: deprecate this macro for naming consistency.
- Builder.defineMacro("__AMDGCN_WAVEFRONT_SIZE", Twine(WavefrontSize));
+ Builder.defineMacro("__AMDGCN_WAVEFRONT_SIZE__", Twine(WavefrontSize),
+ "compile-time-constant access to the wavefront size will
"
+ "be removed in a future release");
+ Builder.defineMacro("__AMDGCN_WAVEFRONT_SIZE", Twine(WavefrontSize),
+ "compile-time-constant access to the wavefront size will
"
+ "be removed in a future release");
----------------
al42and wrote:
The use we currently have in our project is to detect, via CMake's
`try_compile`, whether we have Wave64 or Wave32 targets (or both) to later
choose which data structure flavors to instantiate. With the macro, we can
easily do `#if`+`#warning` inside a kernel, and then parse the build log.
Without it, the only solution we found was to parse the list of targets, which
feels more brittle. `__builtin_amdgcn_wavefrontsize()` is not a compile-time
constant expression, so cannot be used for such checks. Am I missing something?
What is the planned timeline for the removal of this macro?
https://github.com/llvm/llvm-project/pull/112849
_______________________________________________
cfe-commits mailing list
[email protected]
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
