lei137 wrote: > Will this affect the code that is generated for ‘__vector_quad’ values? There > is no direct way to load and store those so they will use pair loads and > stores.
Not AFAIK. The change will get the alignment from the type given. The LIT test being updated also have `alloca` and `store` of vector quads. The alignment for those instructions stayed 64bit. https://github.com/llvm/llvm-project/pull/108606 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
