AMD R700 2D Driver Performance Comparison [Archive]

phoronix

01-23-2009, 03:10 PM

d2kx

01-23-2009, 03:33 PM

To be honest, I would have expected worse results. R7xx seems to run ok for 2D even in this early stage of support.

bridgman

01-23-2009, 04:56 PM

That's better than I expected too. I was a bit worried when I saw the article since the focus has been entirely on getting the basic functions working, not performance optimization of any kind.

stan

01-23-2009, 05:52 PM

It would be interesting to see a comparison of the two open source drivers, radeonhd and radeon on a card that is supported by both.

cruiseoveride

01-23-2009, 06:29 PM

What is the 2D performance of this driver vs the Nvidia proprietary driver?

bulletxt

01-23-2009, 07:26 PM

What is the 2D performance of this driver vs the Nvidia proprietary driver?

Don't ask such questions!

Michael

01-23-2009, 08:46 PM

I was a bit worried when I saw the article since the focus has been entirely on getting the basic functions working, not performance optimization of any kind.

Yeah, I hadn't intended on doing an R700 2D performance look this early, but it was requested by Matthew. It turned out pretty well for a basic implementation :)

bulletxt

01-23-2009, 09:20 PM

Yeah, I hadn't intended on doing an R700 2D performance look this early, but it was requested by Matthew. It turned out pretty well for a basic implementation :)

let's say the truth: what turned out is that AMD should be ashamed of what they call a driver (fglrx). Let's just hope all this sh%$ comes to an end with a complete open source driver one day. And actually yes, do a comparison with NVIDIA proprietary driver if you can... let's see what comes out... you never know.. maybe even NVIDIA drivers suck?

miles

01-24-2009, 12:39 AM

What is the 2D performance of this driver vs the Nvidia proprietary driver?

If you manage to use the Nvidia proprietary driver on the RV770, I'll be really interested by your results.

Darkfire Fox

01-24-2009, 12:52 AM

let's say the truth: what turned out is that AMD should be ashamed of what they call a driver (fglrx). Let's just hope all this sh%$ comes to an end with a complete open source driver one day. And actually yes, do a comparison with NVIDIA proprietary driver if you can... let's see what comes out... you never know.. maybe even NVIDIA drivers suck?

Last I heard, didn't Nvidia's newest drivers produce crummy 2D performance on GeForce 8000 and newer?

In any case, I am totally ecstatic that the EXPERIMENTAL r700 code has already come this far, even surpassing the proprietary Catalyst driver in 2D acceleration. (not that fglx sets the bar very high... )

With any luck, end-users like me will have a working r600/r700 driver when Ubuntu 9.04 comes out. :D

Redeeman

01-24-2009, 04:39 AM

Last I heard, didn't Nvidia's newest drivers produce crummy 2D performance on GeForce 8000 and newer?

In any case, I am totally ecstatic that the EXPERIMENTAL r700 code has already come this far, even surpassing the proprietary Catalyst driver in 2D acceleration. (not that fglx sets the bar very high... )

With any luck, end-users like me will have a working r600/r700 driver when Ubuntu 9.04 comes out. :D
that depends on what "working" requires.. :) if it requires opengl 2.1 and stuff, i think you miiigghhtt wanna switch the timetable a tad :)

this is nice though

d2kx

01-24-2009, 05:26 AM

Yeah, I hadn't intended on doing an R700 2D performance look this early, but it was requested by Matthew. It turned out pretty well for a basic implementation :)

So maybe they decide again on whether they will implement EXA in the fglrx driver. Because this comparison does not show how great RadeonHD is (but it will be, of course), but how slow fglrx is.

yoshi314

01-24-2009, 07:23 AM

i thought radeonhd will get beaten on all tests. instead, it appears that fglrx has still numerous performance issues.

i wonder if it's due to complexity of the code, or just bugs. there was a claim that simpler driver might actually beat more complex driver.

bridgman

01-24-2009, 07:30 AM

My interpretation of the test results was that "EXA is finally worth implementing" ;)

i thought radeonhd will get beaten on all tests. instead, it appears that fglrx has still numerous performance issues. I wonder if it's due to complexity of the code, or just bugs. there was a claim that simpler driver might actually beat more complex driver.

This is actually a real good example of the relative benefits of open vs closed source. 2D is relatively simple in terms of the amount of code required, so a single good developer can produce a sufficiently optimized driver in a practical amount of time. In that case the open source driver will usually win because it can and will track improvements elsewhere in the stack more quickly (since whoever changes the rest of the stack will probably change the driver at the same time).

3D, by comparison, is maybe 100x the code size (seriously) so there the advantages of being able to share proprietary code across a number of OSes will typically outweigh the benefits of having source code available to developers working on other parts of the stack.

Gallium3D is interesting because (a) it seems to have the API lines in the right place to let relatively small HS-specific code still get decent performance, and (b) there is some cross-OS potential so development effort can bleed in from other OSes and markets.

energyman

01-24-2009, 02:53 PM

llama

01-24-2009, 06:05 PM

Phoronix: AMD R700 2D Driver Performance Comparison

Earlier this week we delivered results from a comparison between the Catalyst and X.Org Radeon drivers looking at the R500 2D performance. With a lot of interest having been generated from that, we have now carried out the same set of tests again but this time using an ATI Radeon HD 4850 (RV770) graphics card and the experimental EXA support.

http://www.phoronix.com/vr.php?view=13413

Very interesting results. Your conclusion is a little funny, though: The important thing isn't which driver won more benchmarks. It's which driver had big slowdowns that will be a bottleneck for some real-world activities and e.g. make firefox feel slow.

There were only a few cases where one driver was much slower than the other. IMHO, you should be focusing on those. e.g. Catalyst took 0.09s vs. radeonhd's 0.12s on GtkTextView scrolling. Text scrolling is definitely something that can make a desktop feel slow. e.g. server mobos with onboard mach64 chips suck so much with gnome-terminal that xterm is preferable. (Fortunately, the newer ATI E1000 chipset is up to par and can run a gnome desktop pretty well. No 3D accel, though, since the HW omits that functionality!)

energyman

01-24-2009, 08:10 PM

well, gnome sucks anyway ....

Enverex

01-25-2009, 06:11 AM

that circles test result looks fishy. How many test runs were made? Just one or several (and I don't mean gtkperf rounds...)? Maybe it was some background activity? A ubuntu bug or a bug in the pts? Because 4000 seconds.. that is completly out of the loop.

The fglrx driver sucks on that test and one of the other tests. I tried it myself on my HD4850 and it took a very long time.

Also don't bother with more comments like your last about Gnome, that is just pointless trolling.

tmpdir

01-25-2009, 08:03 AM

well, gnome sucks anyway ....

Than you'll love this article ;)

thread=15059 (http://www.phoronix.com/forums/showthread.php?t=15059)

makomk

01-25-2009, 10:46 AM

Very interesting results. Your conclusion is a little funny, though: The important thing isn't which driver won more benchmarks. It's which driver had big slowdowns that will be a bottleneck for some real-world activities and e.g. make firefox feel slow.

There were only a few cases where one driver was much slower than the other. IMHO, you should be focusing on those. e.g. Catalyst took 0.09s vs. radeonhd's 0.12s on GtkTextView scrolling. Text scrolling is definitely something that can make a desktop feel slow. e.g. server mobos with onboard mach64 chips suck so much with gnome-terminal that xterm is preferable. (Fortunately, the newer ATI E1000 chipset is up to par and can run a gnome desktop pretty well. No 3D accel, though, since the HW omits that functionality!)

I think text scrolling often involves copies with overlapping source and destination, and the current code for that in the radeonhd r6xx-r7xx-support branch is less than ideal. (More specifically, it does the copy a line/column at a time.)

energyman

01-25-2009, 11:32 AM

The fglrx driver sucks on that test and one of the other tests. I tried it myself on my HD4850 and it took a very long time.

Also don't bother with more comments like your last about Gnome, that is just pointless trolling.

well, gnome IS slow. There is no doubt about that. And compiz is pretty broken under its shiny surface so both make a less than ideal platform.

I didn't try 8.12 but a friend with a 48XX and I with my 3870 both tried gtkperf with the leaked 9.2 preview - and circles was fast for both of us.

energyman

01-25-2009, 11:35 AM

Than you'll love this article ;)

thread=15059 (http://www.phoronix.com/forums/showthread.php?t=15059)

have you even read the interview? Linus switched because Fedora - the old KDE haters, forced 4.0 down his throat - in a way that everything was broken afterwards and he was just too lazy to clean that mess up. Not KDE's fault, who told everybody not to use 4.0 - and not because gnome is so great. I am wondering why he is still using broken-OS aka fedora in the first place.

SolidSteel144

01-25-2009, 01:01 PM

Maybe the ATi developers should just contribute to the open source driver instead of their proprietary version.

Enverex

01-25-2009, 01:06 PM

well, gnome IS slow. There is no doubt about that. And compiz is pretty broken under its shiny surface so both make a less than ideal platform.

I didn't try 8.12 but a friend with a 48XX and I with my 3870 both tried gtkperf with the leaked 9.2 preview - and circles was fast for both of us.

Gnome may be slow for -you-, but I've never had any performance issues on any of my machines with any cards.

bridgman

01-25-2009, 01:28 PM

Maybe the ATi developers should just contribute to the open source driver instead of their proprietary version.

The two drivers really serve different markets; it's not really "one or the other". The open driver will normally track new things in the stack (like the recent EXA improvements) more closely than fglrx, but the fglrx driver will have advantages in areas where being able to leverage code across other closed drivers makes the biggest difference, such as 3D performance.

Enverex

01-25-2009, 01:30 PM

That sounds almost like saying "Well you can have this driver and have good 2D performance, or have this driver and have good 3D performance". Doesn't sound ideal.

bridgman

01-25-2009, 01:56 PM

This is pretty much a one-time situation. It just so happened that EXA improvements were the most recent thing to show up in the released framework and the open driver can take advantage of them first.

Most of the upcoming framework improvements (memory management, Gallium etc..) will be complete non-issues since "all" they do is allow open drivers to get closer to closed drivers in terms of performance and features. I'm not trying to downplay them -- I think they are hugely important for the X community -- but the closed drivers have had similar capabilities for years.

cruiseoveride

01-25-2009, 03:14 PM

Maybe the ATi developers should just contribute to the open source driver instead of their proprietary version.

Hope to God they dont. I shit driver is enough.

SolidSteel144

01-25-2009, 03:18 PM

Hope to God they dont. I shit driver is enough.

Oh come on, I'm sure they would provide some help.

energyman

01-25-2009, 03:21 PM

they do already

rampage7

01-28-2009, 02:50 PM

I've installed this driver on my Gentoo system with rv770 4850 graphic card. Performance in KDE4 apps (dolphin, plasma) is great, but only witch enabled composite (by running xcompmgr). Without composite enabled, moving windows is very slow.

But I've noticed a problem - i can't enable XRender-based desktop effects in kwin configuration window - it displays standard dialog without any usable information (it says that activating windows effects failed). Xcompmgr works without any problems. Any ideas how to force kwin to worh with experimental r600/700 radeonhd driver? I've tried "disable functionality check" option without any success.

energyman

01-28-2009, 03:04 PM

your driver installation is borked.

rampage7

01-28-2009, 03:15 PM

your driver installation is borked.

it's not helping me or anyone.

I know that exa acceleration is working (very fast with xcompmgr running). This is the only usable functionality of radeonhd driver from r600/700 support branch. So, if exa is working, my installation is far from broken.

Anyone tried using this driver with kwin from kde4?

energyman

01-28-2009, 03:21 PM

oh, you are talking radeonhd ....

ech0s7

01-30-2009, 05:22 AM

if you would see my comparison test, this is the link: http://www.ech0s7.netsons.org/index.php/archives/31

ech0s7

energyman

01-30-2009, 09:17 AM

ech0s7

01-30-2009, 10:18 AM

are you use that the tests are running correctly?
Because when run with effects turned on gtkperf seems to run faster - but it also looks like some stuff simple never got drawn. So maybe radeonhd is much faster because it does not do all the stuff it is supposed to do?

Yes, i run correctly the test, and without any effect. GtkPerf not seems to run faster, it run faster! In any case, the difference is remarkable and impressive :eek: !