[Bug target/85323] SSE/AVX/AVX512 shift by 0 not optimized away

kretz at kde dot org gcc-bugzilla@gcc.gnu.org
Tue Apr 10 13:08:00 GMT 2018


https://gcc.gnu.org/bugzilla/show_bug.cgi?id=85323

--- Comment #1 from Matthias Kretz <kretz at kde dot org> ---
Created attachment 43898
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=43898&action=edit
idea for a partial solution

Constant propagation works using the built in shift operators. At least for the
shifts taking a "const int" shift argument, this is a possible fix. However,
the patched intrinsics now accept non-immediate shift arguments, which modifies
the interface (in an unsafe way).


More information about the Gcc-bugs mailing list