ATI Radeon X800 XT Mac Edition: Competition in the Mac Market
by Anand Lal Shimpi on January 5, 2005 7:38 PM EST- Posted in
- Mac
UI Performance
Mac OS X 10.3.7 relies on a fully OpenGL accelerated GUI to, in a sense, make things look pretty (and enable things like Exposé). But despite what you may think, the majority of UI performance is still dependent upon the CPU. Until the release of OS X Tiger all UI elements are CPU rendered and with the appropriate GPU are simply treated as AGP textures and composed on the GPU itself. Although the majority of the work is still done on the CPU, there is an element of GPU interaction that can impact performance.
In order to measure the GPU’s impact on overall UI responsiveness we turn to XBench, a synthetic test that can give us a slight idea of UI performance. The three tests we focus on are XBench’s Quartz, OpenGL and UI tests. The latter is possibly the most important to UI performance but all are interesting.
The Quartz test focuses on Quartz rendering performance and is thus mostly CPU bound, but as we mentioned earlier there is a slight impact of graphics performance. More than anything we’re looking at driver maturity here, since the test uses almost 100% of the CPU (it is only a single threaded test).
The X800 XT Mac Edition offers a negligible performance improvement over the 9800 of just under 4%. The 6800 Ultra is a bit faster at 9%, but compared to the X800 it’s only 5% faster. When looking at real world performance impacts we’re talking about wholly unnoticeable however. It is a good thing to see that NVIDIA’s Mac drivers are doing well at this stage, which was an issue when the 6800 was first launched for the Mac platform.
The OpenGL performance test is more of a primitive 3D test than anything else, it doesn’t really impact UI performance at all – it’s more of a basic texturing test than anything else as it makes no use of complex shaders.
It’s no surprise that the ATI cards do better as ATI is stronger at basic texturing performance.
The most interesting test is the UI test, which basically tests multiple UI elements and scrolling performance under OS X.
Here the X800 was actually slightly slower than the 9800, but most importantly the 6800 Ultra actually ended up being about 9% faster than the X800, which is a pretty decent margin. We’ve asked ATI about general UI performance under OS X Panther however we’ve yet to receive a response as to why it feels slow in certain situations, especially with smooth scrolling enabled. While this will most likely be a moot point once Tiger ships (as pixel shader performance will matter much more than AGP texturing performance), it is an issue today and there’s no excuse for ATI to be slower in this test.
With the synthetic benchmarks out of the way, let’s look at some games in order of decreasing importance.
Facebook
Twitter
RSS
35 Comments
View All Comments
gizzard - Saturday, January 8, 2005 - link
Hey, that's me! =)Quite simply, the reason the ATI card is faster is because it has better drivers, (The higher core clock speed helps also.) You can really see it in the lower resolution Quake3 scores when even the 9600 XT and 9800 Pro beat out the 6800. Driver inefficiencies can sometimes be exposed by looking at low resolution performance, which is why I always publish them.
It is difficult to see the large performance deltas in Anand's review because he didn't perform many tests with FSAA and AF enabled — that is where the X800 XT shines. Also, his conclusion of the RtCW scores was not entirely accurate: "The story is pretty simple here, on older games, the 9800 is already CPU bound thus making the 6800 Ultra and X800 XT not too interesting to look at." RtCW is a heavily CPU bound game. Given the same CPU, you won't see a large delta unless you compare a Rage 128 to an X800 XT — and his results show that. I'm just not sure how he arrived at that conclusion . =)
I'm hoping that Anand will speak up in the comments section to clear things up. I'd be very interested in what he has to say. And as everyone said, what Anand is doing is A Good Thing.™
gankaku - Saturday, January 8, 2005 - link
I was interested to read Anand's take on the new ATi X800 XT for several reasons, including:1) his earlier fair, well-written article on a month with a Mac
2) his status in the PC World as an honest broker, and his rigorous testing procedures
3) the fact that his opinion does carry some weight, and may help improve hardware offerings in the Mac world. We've got it pretty good right now, but articles that spread light (and not just heat) are always appreciated.
But, after all that, his results surprised me. That's because I read an earlier review of the new ATi card at Inside Mac Games, with the results indicating that the ATi card is a much better buy.
I'm not enough of a geek to understand why ATi scored better at IMG, but I'm sure that some lively debate on this list will release both heat and light in the subsequent discussions here.
http://www.insidemacgames.com/reviews/view.php?ID=...
karlreading - Friday, January 7, 2005 - link
personally, i like macs about as much as i like bineg shot in my stomache and left to bleed to death. that said, all these people dissing anad for relesing a mac article are very silly. Its very refreshing to see a pc orientated site like anadtech give readers exposer to other computer circles, and whilst it exposes pc users to the mac side of things, it also may help bring some mac users to anandtech, and anything that brings more COMPUTER users together is welecomed IMHO.keep up the good work anand!!
a2daj - Friday, January 7, 2005 - link
Forgot to add that 10.3.7 added some major shader performance improvements for Radeon 9600s and up, which in turn helped Halo performance.a2daj - Friday, January 7, 2005 - link
As for Halo and FSAA:Mac Halo has always been able to do FSAA, but various driver bugs would effect different cards, essentially breaking the feature for some users. And yes, it looks great with FSAA ;) But most of those bugs have been worked out in various OS and Halo patches. Mac Halo 1.5 introduced hardware accelerated Lens Flares for users with OS 10.3.5 (presumably through the use of ARB Occlusion Query, which was finally implemented in Apple's OpenGL implementation with the release of 10.3.5). The previous combination of Halo Lens Flares (1.02-1.05.3) and FSAA could cause major performance hits, particularly in the last level during the escape. The older versions of Halo had to rely on glReadPixel to handle Lens Flares, which resulted in noticeable performance hits with Lens Flares set to High and Extreme. The Performance hit was made even worse when FSAA was enabled. Halo 1.5 fixed that.
a2daj - Friday, January 7, 2005 - link
For the G5 cards which need more power, the PRO in AGP Pro is supplying the extra power for the cards. That's the whole point of the PRO in AGP Pro slots. The ADC Power Nubin is just for ADC.aliasfox - Thursday, January 6, 2005 - link
Frame rates are lower than the PC scores for two reasons:- Most games outside of games based off of the Doom 3 and Quake III engines are DirectX, something that is nowhere on the Mac. So not only are all of these games ported over to a different processor architecture, but also over to OpenGL.
- The graphics cards were also tested on a 2.0 GHz G5- while still fast, is slower than a more modern 2.5 GHz machine. The 2.0 GHz was originally released in May 2003- roughly the same time as the FX-51 and 3.0 GHz P4. So yes, these scores are lower than those produced by top end PCs because it's older (though still not necessarily cheaper).
Chuckles - Thursday, January 6, 2005 - link
#23:A big chunk of the additional power for graphics cards comes from the 25-28V rail in G4's and G5's. This additional power is normally run though additional molexes in PC's, but in a Mac, the power goes through a seperate set of pins on the external end of the AGP (PRO) slot (the two-pin tab between the end of the DVI and AGP PRO connector).
Incidentally, on older Apple displays, the display itself was powered through the computer's 25V rail. This is no longer the case partially due to the 30" display's 150 W draw (6A @ 25V).
On another note. It's kind of annoying to see things CPU bound at high resolution.
mbhame - Thursday, January 6, 2005 - link
OriginalReaper - is that YOU???CindyRodriguez - Thursday, January 6, 2005 - link
Just finished...In my humble opinion, a much improved article to add to your Mac section.
Again, thanks for the Mac article done in a fair way. It's understandable difficult to bench video well on a Mac.
I'd be interesting to have seen a comparison of stock mac configs, like an apple branded 9600xt or the stock 5200 Ultra.
I'd also be interesting to see a comparison of a similar PC running an x800 with the same benchmarks... that is interesting from the perspective of how a competative market affects driver optimization efforts.
cheers,
cindy