[llvm] VectorWiden pass to widen aleady vectorized instrctions (PR #67029)

Dinar Temirbulatov via llvm-commits llvm-commits at lists.llvm.org
Thu Sep 28 23:13:58 PDT 2023


dtemirbulatov wrote:

> Have you looked at doing this in VectorCombine?
Yes, VecotorCombine deals with a limited number of operations and ussually dependent. Still, we want to widen the whole chain of operations independet from each other, from loads to stores, to build something like an SLP tree.


https://github.com/llvm/llvm-project/pull/67029


More information about the llvm-commits mailing list