This is the mail archive of the gcc-bugs@gcc.gnu.org mailing list for the GCC project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[Bug rtl-optimization/46391] false dependencies are computed after vectorization (#2)


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=46391

--- Comment #1 from Richard Guenther <rguenth at gcc dot gnu.org> 2010-11-09 13:18:47 UTC ---
On x86_64 restrict information is there (-fdump-tree-optimized-alias):

<bb 5>:
  # PT = nonlocal { PARM_RESTRICT.0 } (restr)
  vect_p.27_71 = c_8(D) + prolog_loop_niters.16_27;
  # PT = nonlocal { PARM_RESTRICT.1 } (restr)
  vect_p.32_75 = d_5(D) + prolog_loop_niters.16_27;

<bb 6>:
  # ivtmp.52_46 = PHI <ivtmp.52_31(6), 0(5)>
  # ivtmp.55_47 = PHI <ivtmp.55_33(6), 0(5)>
  vect_var_.28_74 = MEM[base: vect_p.27_71, index: ivtmp.55_47, offset: 0B];
  MEM[base: vect_p.32_75, index: ivtmp.55_47, offset: 0B] = vect_var_.28_74;
  ivtmp.52_31 = ivtmp.52_46 + 1;
  ivtmp.55_33 = ivtmp.55_47 + 16;
  if (ivtmp.52_31 < bnd.20_49)
    goto <bb 6>;
  else
    goto <bb 7>;


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]