Hide Forgot
AVX2 is already supported in Broadwell and Haswell, we should be able to test it. Note that avx2 support was added, not avx-512. So maybe the summary should be updated?
(In reply to Amit Shah from comment #2) > AVX2 is already supported in Broadwell and Haswell, we should be able to > test it. > > Note that avx2 support was added, not avx-512. So maybe the summary should > be updated? Yep, you're right, so we should be able to test it on an AVX2 machine which we'll have somewhere.
(In reply to Dr. David Alan Gilbert from comment #3) > (In reply to Amit Shah from comment #2) > > AVX2 is already supported in Broadwell and Haswell, we should be able to > > test it. > > > > Note that avx2 support was added, not avx-512. So maybe the summary should > > be updated? > > Yep, you're right, so we should be able to test it on an AVX2 machine which > we'll have somewhere. "grep avx2 /proc/cpuinfo" will confirm if the machine has the extensions which this feature uses.
OK, it looks like we have two problems: a) the configure time test gets optimised out with -O2 and decides it can't do AVX2 b) There's a compile time test in the code that uses it which checks that it has gcc4.9 or better to work around a header bug, but we've only get 4.8.5 in RHEL. Paolo says that's due to https://gcc.gnu.org/ml/gcc-patches/2013-06/txt6Tt74laQsp.txt which we don't seem to have yet.
Fix included in qemu-kvm-rhev-2.6.0-8.el7
Move to verified per comment 13 & 14
Since the problem described in this bug report should be resolved in a recent advisory, it has been closed with a resolution of ERRATA. For information on the advisory, and where to find the updated files, follow the link below. If the solution does not work for you, open a new bug report. https://rhn.redhat.com/errata/RHBA-2016-2673.html