[Bug tree-optimization/103144] vectorizer failed to recognize shift>>=1 in loop as shift>>i

[rguenth at gcc dot gnu.org via Gcc-bugs]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103144

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2021-11-09
             Blocks|                            |53947
             Status|UNCONFIRMED                 |NEW
                 CC|                            |rguenth at gcc dot gnu.org
     Ever confirmed|0                           |1

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
There are two choices.

1) implement more general induction support and vectorize it as

    shiftv = { shift, shift >>1, shift>>2, shift>>3 };
    for (int64_t i = 0; i != 64; i++)
    {
      pdst[i] = psrc[i] + shiftv;
      shiftv >>= {4, 4, 4, 4};
    }

(I think I've seen related bugreports doing bit-flip "induction" with XOR)

2) add the ability to have patterns for PHIs and do what you suggest,
pattern-replacing the shift PHI def with a non-PHI pattern shift >> i
after doing cycle analysis - then the PHI and the shift >>= 1 should become
!STMT_VINFO_RELEVANT.


Referenced Bugs:

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=53947
[Bug 53947] [meta-bug] vectorizer missed-optimizations