http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51499
--- Comment #14 from Ira Rosen <irar at il dot ibm.com> 2011-12-13 16:27:19 UTC --- (In reply to comment #13) > > However, I don't fully understand Richard Guenther's example. Yes his > example requires -fassociative-math to be vectorized, however, my example > would translate to something like > > sum1 += a[i]; > sum2 += a[i+1]; > > and now it doesn't matter if it's executed this way or the other way > around > > sum2 += a[i+1]; > sum1 += a[i]; The problem is probably more in implementation. The change of order will also change between sum1 and sum2, so when you want to return sum1+sum2, the vectorized version will return sum2+sum1.
