================ @@ -4899,6 +4899,238 @@ void CGOpenMPRuntime::emitSingleReductionCombiner(CodeGenFunction &CGF, } } +void CGOpenMPRuntime::emitPrivateReduction( ---------------- chandraghale wrote:
- Updated the code with comments and clarified the logic - Updated OpenMPSupport.rst and the release notes. - Unified the logic as much as possible. - In regular reduction codegen, threads update a shared variable directly. With private reduction, each thread first works with its own private copy, and then all these partial results are combined into a shared variable. The final result is copied back to each thread’s original variable https://github.com/llvm/llvm-project/pull/134709 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
