https://gcc.gnu.org/bugzilla/show_bug.cgi?id=83064
Dominique d'Humieres <dominiq at lps dot ens.fr> changed: What |Removed |Added ---------------------------------------------------------------------------- CC| |rguenth at gcc dot gnu.org --- Comment #3 from Dominique d'Humieres <dominiq at lps dot ens.fr> --- I confirm the results in comment 0 with revisions r254869 and r255053, except I get 'always zero' with 'unrolled do-loop' and '-O1'. With revision r254940, the 'unrolled do-loop' is parallelized with '-O1', '-O2', and '-Ofast' only, with a wrong result for '-O1'. The other tests are parallelized with '-O1' and '-O2' only, with a wrong result in both cases. For all revisions the 'unrolled do-loop' is parallelized with '-ftree-parallelize-loops=4 -Ofast' only, with the right result.