[Bug target/88152] optimize SSE & AVX char compares with subsequent movmskb

[glisse at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=88152

Marc Glisse <glisse at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2018-11-22
     Ever confirmed|0                           |1

--- Comment #1 from Marc Glisse <glisse at gcc dot gnu.org> ---
Turning all the comparisons into a<0 is something we do already at the GIMPLE
level for scalars. Some more restricted form should extend to vectors.

For the combination movmsk(x<0), we could add such an insn to sse.md, which
would generate only movmsk, or we could transform movmsk(x<0) into movmsk(x).