You are not logged in.
Pages: 1
I just got sent this link: http://tech.slashdot.org/article.pl?sid … 13/0058251
Apparently, the windows version of Firefox runs better under wine than the native linux version, possibly because the linux version isn't optimized. Can anyone take a look at this and see if they can confirm it?
Offline
I confirmed it this morning. It's significant, even on Gentoo, where I've compiled it specifically to a particular CPU. Hopefully they'll look at some optimization on it, but I'm not really worried. Firefox is plenty responsive for my uses, so I'll keep using it.
-mS
Last edited by aglarond (2009-02-13 22:36:38)
Offline
The case is probably slow and old gtk, not so good cairo support etc. Firefox interface is dog slow on my PC. Thankfully they're porting Firefox to QT (second time...).
Offline
No. The tests that were used don't do any graphical rendering (try doing this test and see for yourself). Incidentally, if you read the comments made by the author of the qt port of firefox you'll see that gtk/cairo based rendering is still, at this stage at least, faster and more efficient than what the qt version is capable of.
Last edited by fwojciec (2009-02-24 12:01:50)
Offline
No. The tests that were used don't do any graphical rendering (try doing this test and see for yourself). Incidentally, if you read the comments made by the author of the qt port of firefox you'll see that gtk/cairo based rendering is still, at this stage at least, faster and more efficient than what the qt version is capable of.
Did the test on both here (arch-64, firefox-pgo 3.06) and at work (windows xp, firefox 3.06)
windows:
Total: 5723.8ms +/- 7.9%
arch:
Total: 2479.6ms +/- 0.9%
firefox-pgo from aur
"I know what you're thinking, 'cause right now I'm thinking the same thing. Actually, I've been thinking it ever since I got here:
Why oh why didn't I take the BLUE pill?"
Offline
Did the test on both here (arch-64, firefox-pgo 3.06) and at work (windows xp, firefox 3.06)
windows:
Total: 5723.8ms +/- 7.9%
arch:
Total: 2479.6ms +/- 0.9%firefox-pgo from aur
Different PC's, different architectures, I don't see how these results have any meaning. I think the original poster was saying that on the SAME computer, the firefox natively compiled for linux was slower than running the windows version of firefox in WINE.
Cheers,
Alphalutra1
Offline
fwojciec wrote:No. The tests that were used don't do any graphical rendering (try doing this test and see for yourself). Incidentally, if you read the comments made by the author of the qt port of firefox you'll see that gtk/cairo based rendering is still, at this stage at least, faster and more efficient than what the qt version is capable of.
Did the test on both here (arch-64, firefox-pgo 3.06) and at work (windows xp, firefox 3.06)
windows:
Total: 5723.8ms +/- 7.9%
arch:
Total: 2479.6ms +/- 0.9%firefox-pgo from aur
I checked on a single machine, with regular ff in windows and a pgo version in linux: firefox in windows is about 10-15% faster.
Last edited by fwojciec (2009-02-25 06:29:37)
Offline
Back when that Slashdot article appeared I tried the nightly builds of 3.1 back-to-back on this machine. There's no doubt the Windows version under WINE comes out on top in the benchmarks. But using it for five minutes normal browsing convinced me you're better off sticking to a native build. Whether it's WINE or what, I don't know, but in normal usage it feels a lot slower.
0 Ok, 0:1
Offline
I did the test fwojciec mentioned in post #4. I have tested Firefox under Linux and Windows, and there's no difference.
Offline
I just did it on arch and got
5739.4ms; haven't tried windows
Any society that would give up a little liberty to gain a little security will deserve neither and lose both.
-Benjamin Franklin
The reasonable man adapts himself to the world; the unreasonable one persists in trying to adapt the world to himself. Therefore all progress depends on the unreasonable man.
-George Bernard Shaw
Offline
Yeah, there is no difference between the same machine running windows and firefox than linux and firefox. The difference is only when it is run on wine, so this is even better for the wine developers.
Last edited by kensai (2009-02-24 22:27:26)
Offline
Yeah, there is no difference between the same machine running windows and firefox than linux and firefox. The difference is only when it is run on wine, so this is even better for the wine developers.
That is definitely not true in my tests. I've done it multiple times, because I though that openjdk6 and jde/jre could make a difference, and I always got results that were consistent with the findings of the Slashdot article. I tested on a fairly slow computer (Intel Atom based netbook) and the performance differences were very clear. I'm not sure why others are getting different results -- perhaps pgo-optimized builds of Firefox are slower than the generic Arch build? I'm very skeptical of those findings that claim that performance is the same in Linux and Windows, because Linux/Windows performance actually differs for every individual test in sunspider test suite -- in some cases Linux is actually faster, but in most cases Windows versions of Firefox performs better.
Last edited by fwojciec (2009-02-25 06:32:43)
Offline
--NOTE : I didn't read the story comments--
I wonder If the Firefox guys use gcc to build for Windows (and what parameters they pass to the compiler) .
An interesting test case is to build Firefox for both Linux and Windows with similar features and build configurations .
English is not my native language .
Offline
I don't see any difference between
arch firefox 3.06 and arch firefox-pgo 3.06 from aur
Both generate exactly the same results:
http://www2.webkit.org/perf/sunspider-0 … pider.html
arch firefox:
Total: 4490.0ms +/- 1.4%
arch firefox-pgo
Total: 4462.6ms +/- 0.9%
http://v8.googlecode.com/svn/data/bench … 3/run.html
arch firefox 3.06
130
arch firefox-pgo
127
In other words there is much ado about nothing.
I am using linux since 1995 but I have yet to see desktop app optimized for linux (distro) and the same app optimized for windows xp which would run faster on linux.
Comparing 32-bit xp with 64-bit linux (as someone did) was a joke?
Offline
It's not about pgo/generic. It's about Firefox performance in Linux and Windows/Wine.
Offline
I know, but someone was boldly testing firefox-pgo (on slashdot pgo argument was even used as a major critique -> windows pgo and linux generic). So I tested both and pgo performs no better that generic.
What I am saying is that both versions of arch/linux firefox are slow (as all desktop apps under linux).
To summarize pgo does nothing to spped up firefox under linux and firefox is slower under linux that xp.
Offline
Sorry, I misunderstood what you meant...
Offline
no problem (I was not clear).
in fact I was hoping that pgo would help to some extent..
maybe next time.
Offline
I am using linux since 1995 but I have yet to see desktop app optimized for linux (distro) and the same app optimized for windows xp which would run faster on linux.
I don't know if it's true anymore, but Blender used to at every level provided that the graphics driver was nVidia. That's a whole different beast though, and I don't think that it's a fair comparison in terms of a 'desktop' app.
Offline
Pages: 1