[PATCH] D34472: [DAG] Rewrite areNonVolatileConsecutiveLoads to use BaseIndexOffset

[Simon Pilgrim via Phabricator via llvm-commits]

RKSimon added a comment.

> - test/CodeGen/X86/clear_upper_vector_element_bits.ll - DAG is improved, but scheduling causes a code size degradation. This is probably okay.

That's alright - that test is a rather convoluted case, at some point the build vector lowering should be able to avoid it properly.


https://reviews.llvm.org/D34472