| View previous topic :: View next topic |
| Author |
Message |
Tony0945
Watchman
Joined: 25 Jul 2006
Posts: 5127
Location: Illinois, USA
|
 Posted: Thu Dec 21, 2017 3:32 am Post subject: GCC timing changes after switching from13.0 profile to 17.0 |
 |
|
About a week ago I transitioned a desktop system from 13.0 to 17.0
I had promised to do before and after gcc timings and post the results.
The system was up to date and running gcc 6.4.0 for some time. All software had been rebuilt with gcc 6.4.0 when moving from gcc 6.3.0 to gcc 6.4.0 therefore emerge -e world was not run as part of this update. The update was accomplished by
| Code: | eselect profile {number of 17.0/Desktop}
. /etc/profile
emerge -1v gcc
emerge -1v sys-devel/libtool
emerge -1v binutils
emerge -1v glibc
emerge -1a /lib*/**.a /usr/lib*/**.a
emerge -auvND world |
Tests were run before and after switching to 17.0
All 13.0 tests were run with gcc-6.4.0 -pie.
All 17.0 tests were run with gcc-6.4.0 +pie
All tests were run with the following services stopped: xdm, samba, minidlna, dbus. Openssh was running but no one was logged in.
Times were taken with genlop -t. These times were 5 to ten seconds slower than those measured by “time”. Each package was emerged with “time emerge -1 <packagename> --nodeps”.
--nodeps was used to avoid measuring time for portage to examine the dependency tree. all deps were known to be satisfied.
| Code: | Package 13.0 compile time 17.0 compile time. extra for 17.0
gcc 1h 3m 9s 1h 2m 30s -39 seconds 1.04% faster
firefox 42m 39s 42m 42s 3 seconds 0.12% slower
wxGTK 7m 59s 7m 59s 0 seconds no change
abiword 9m 47s 10m 12s 25 seconds 4.26% slower
mate-control-center 1m 39s 1m 39s 0 seconds no change |
System specs : Athlon X3 4G RAM 10,000 RPM hard drive. 1G swap partition plus 12G swapfile running Mate 1.12 (stopped for test) and OpenRC with eudev. |
|
| Back to top |
|
 |
Ant P.
Watchman
Joined: 18 Apr 2009
Posts: 6920
|
| Posted: Thu Dec 21, 2017 6:25 am Post subject: |
|
|
| Thanks; good to know there's no serious regressions. Maybe there's slight performance differences at runtime, but that takes a lot more effort to measure. Though I'll say I haven't had any visible slowdown since the switch. |
|
| Back to top |
|
|
|
|
You cannot post new topics in this forum
You cannot reply to topics in this forum
You cannot edit your posts in this forum
You cannot delete your posts in this forum
You cannot vote in polls in this forum
|
|
Gentoo Chat
