Usman Nadeem b93173185d [InstCombine] Narrow trunc(lshr) in more cases (#139645) ...

We can narrow `trunc(lshr(i32)) to i8` to `trunc(lshr(i16)) to i8` even
when the bits that we are shifting in are not zero, in the cases where
the MSBs of the shifted value don't actually matter and actually end up
being truncated away.

This kind of narrowing does not remove the trunc but can help the
vectorizer generate better code in a smaller type.
Motivation: libyuv, functions like ARGBToUV444Row_C().

Proof: https://alive2.llvm.org/ce/z/9Ao2aJ

2025-05-13 13:42:42 -07:00

116 KiB

Raw Blame History

View Raw