View previous topic :: View next topic |
Author |
Message |
billydv l33t
Joined: 22 Dec 2006 Posts: 911 Location: Mount Vernon, NY
|
Posted: Wed Apr 24, 2013 1:22 am Post subject: Could Gentoo be slower than Fedora? |
|
|
I have numerous times checked a particular test (peacekeeper browser rating http://peacekeeper.futuremark.com/) to see firefox performance in Gentoo. My Gentoo box is a Maximum PC Dream Machine vs my Fedora box yet consistently my Fedora box beats my Gentoo box by several hundred points. Just tonight 2874 on Gentoo vs 3304 on Fedora on considerably lesser PC hardware. How can this be? _________________ Billy DeVincentis |
|
Back to top |
|
|
PaulBredbury Watchman
Joined: 14 Jul 2005 Posts: 7310
|
Posted: Wed Apr 24, 2013 1:55 am Post subject: |
|
|
Looks like Fedora uses gcc 4.8 already.
Edit: Also, Fedora's xulrunner gets compiled with -Os instead of -O3 (see rhbz-928353.patch, from weird bug). |
|
Back to top |
|
|
ppurka Advocate
Joined: 26 Dec 2004 Posts: 3256
|
Posted: Wed Apr 24, 2013 2:29 am Post subject: |
|
|
PaulBredbury wrote: | Looks like Fedora uses gcc 4.8 already.
Edit: Also, Fedora's xulrunner gets compiled with -Os instead of -O3 (see rhbz-928353.patch, from weird bug). | Bigger ricers than gentoo. _________________ emerge --quiet redefined | E17 vids: I, II | Now using kde5 | e is unstable :-/ |
|
Back to top |
|
|
Ion Silverbolt Apprentice
Joined: 04 Nov 2004 Posts: 203
|
Posted: Wed Apr 24, 2013 4:39 pm Post subject: |
|
|
There are lots of variables we don't know here. What are your cflags in make.conf set to? Which use flags did you enable when you compiled firefox? Which version of gcc did you use? Did you use lto? Many factors can affect performance.
Benchmarking isn't always a reliable reflection of real world performance either. |
|
Back to top |
|
|
Ant P. Watchman
Joined: 18 Apr 2009 Posts: 6920
|
Posted: Wed Apr 24, 2013 6:46 pm Post subject: |
|
|
It's worth remembering that not many Gentoo users have the necessary RAM to build Firefox with the crazy optimisations prepackaged versions use. Mozilla itself is pushing up against the limits of its 32-bit build nodes as of late. |
|
Back to top |
|
|
krinn Watchman
Joined: 02 May 2003 Posts: 7470
|
Posted: Wed Apr 24, 2013 7:50 pm Post subject: |
|
|
Could only answer yes to your question.
If the general consensus (i'm not saying it's true or false) says gentoo can be less bloated than distro XXX, nothing stop someone to bloat his gentoo to a level other distro could hardly reach.
The freedom to have the most ugly, sluggish and slowest distro out there is freedom. |
|
Back to top |
|
|
asturm Developer
Joined: 05 Apr 2007 Posts: 8936
|
Posted: Wed Apr 24, 2013 8:23 pm Post subject: |
|
|
In other words, my Gentoo != your Gentoo |
|
Back to top |
|
|
NeddySeagoon Administrator
Joined: 05 Jul 2003 Posts: 54234 Location: 56N 3W
|
Posted: Wed Apr 24, 2013 8:53 pm Post subject: |
|
|
billydv,
Yep, if you build your Gentoo with -march=i686 it will run on a PentiumPro.
It will not use mmx or any of the other multimedia speedups found on modern CPUs.
It won't even play 1080p HD videos smoothly on your hardware.
You can make it slow if you want to, that the beauty of Gentoos flexibility. _________________ Regards,
NeddySeagoon
Computer users fall into two groups:-
those that do backups
those that have never had a hard drive fail. |
|
Back to top |
|
|
aCOSwt Bodhisattva
Joined: 19 Oct 2007 Posts: 2537 Location: Hilbert space
|
Posted: Wed Apr 24, 2013 9:19 pm Post subject: |
|
|
PaulBredbury wrote: | Looks like Fedora uses gcc 4.8 already.
Edit: Also, Fedora's xulrunner gets compiled with -Os instead of -O3 (see rhbz-928353.patch, from weird bug). |
I also guess that billydv's fedora is F18, in other words a 3.6 kernel... that is... the most efficient money can buy. _________________
|
|
Back to top |
|
|
Yamakuzure Advocate
Joined: 21 Jun 2006 Posts: 2284 Location: Adendorf, Germany
|
Posted: Thu Apr 25, 2013 6:26 am Post subject: |
|
|
I found this interesting, so I made some tests: (I have a laptop with Intel i7 Dual Core, first generation, and 4GB RAM)
First run, firefox-20.0.1 and qupzilla-1.4.1 merged with Code: | CFLAGS="-march=native -O2 -pipe" | and default USE flags:- qupzilla score: 1,762
- firefox score: 2,160
Then I re-emerged only the browsers (qupzilla depends heavily on qtwebkit, which I did not re-emerge!) with Code: | CFLAGS="-march=native -pipe -Os \
-fgraphite-identity -floop-interchange -floop-strip-mine -floop-block \
-ftree-loop-distribution \
-floop-parallelize-all -ftree-parallelize-loops=8" | and added Code: | www-client/firefox custom-cflags system-jpeg system-sqlite | to /etc/portage/package.use:- qupzilla score: 1,841
- firefox score: 2,239
Although it is not that much, it is a definite raise.
However, I didn't feel bold enough to activate USE="custom-optimization" for firefox... Maybe that is what gives the edge. And who knows what gcc-4.8 with lto can do compared to my gcc-4.7.2 without lto... _________________ Important German:- "Aha" - German reaction to pretend that you are really interested while giving no f*ck.
- "Tja" - German reaction to the apocalypse, nuclear war, an alien invasion or no bread in the house.
|
|
Back to top |
|
|
trismo n00b
Joined: 26 May 2012 Posts: 49
|
Posted: Thu Apr 25, 2013 8:47 pm Post subject: |
|
|
this benchmark hard depend on the performance from vpx cairo pango gstreamer not firefox / xulrunner |
|
Back to top |
|
|
Yamakuzure Advocate
Joined: 21 Jun 2006 Posts: 2284 Location: Adendorf, Germany
|
Posted: Mon Apr 29, 2013 9:43 am Post subject: |
|
|
trismo wrote: | this benchmark hard depend on the performance from vpx cairo pango gstreamer not firefox / xulrunner | Well, I did feel bold enough to try out firefox with USE="custom-optimization" and the result dropped from the former 1,841 points down to 1,152. So I daresay the way firefox is built *does* has at least some impact. _________________ Important German:- "Aha" - German reaction to pretend that you are really interested while giving no f*ck.
- "Tja" - German reaction to the apocalypse, nuclear war, an alien invasion or no bread in the house.
|
|
Back to top |
|
|
|