[Bug target/79491] Possibly inefficient code for the inner product of two vectors

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=79491

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |missed-optimization
             Target|                            |x86_64-*-*, i?86-*-*
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2017-02-14
          Component|c                           |target
     Ever confirmed|0                           |1

--- Comment #3 from Richard Biener <rguenth at gcc dot gnu.org> ---
I think that on most archs the cost of unaligned _loads_ are not too bad but
aligning stores can still free up store bandwith.  Unfortunately x86 globs
unaligned load/store into one parameter in its cost tables:

      case unaligned_load:
      case unaligned_store:
        return ix86_cost->vec_unalign_load_cost;

thus confirmed as x86 specific cost model issue.