FYI.
Please participate in the discussion on the devel list, as you see fit.
Thanks,
Florian
-------- Forwarded Message --------
Subject: Compile Fedora with auto-vectorization
Date: Mon, 5 Mar 2018 12:07:40 +0100
From: Florian Weimer <fweimer(a)redhat.com>
Reply-To: Development discussions related to Fedora
<devel(a)lists.fedoraproject.org>
To: Development discussions related to Fedora
<devel(a)lists.fedoraproject.org>
Should we start compiling Fedora with auto-vectorization, either using
-O3 or -O2 -ftree-loop-vectorize?
Downstream experimented with that for POWER 7 (the ppc64p7 packages).
But if auto-vectorization is beneficial on POWER, it likely helps on
other CPUs as well, given that SIMD support is quite common nowadays.
I really want to avoid a package-specific flag because I don't think
discussions about per-package build policies are a good use of our time,
and dependency changes invalidate previous decisions based on package
use all the time. (We saw a lot of that when we still had
package-specific hardening flags.)
We currently have the RPM macro _performance_build for that: It's used
in some Fedora spec files, but the redhat-rpm-config support have never
be contributed to Fedora. If we enable auto-vectorization globally,
those packages could remove the _performance_build setting, a minor cleanup.
Thanks,
Florian
_______________________________________________
devel mailing list -- devel(a)lists.fedoraproject.org
To unsubscribe send an email to devel-leave(a)lists.fedoraproject.org