ATI's X1000 Series: Extended Performance Testing
by Derek Wilson on October 7, 2005 10:15 AM EST- Posted in
- GPUs
Day of Defeat: Source Performance
Rather than test Half-Life 2 once again, we decided to use a game that pushed the engine further. The reworked Day of Defeat: Source takes 4 levels from the original game and adds not only the same brush up given to Counter Strike: Source, but HDR capabilities as well. The new Source engine supports games with and without HDR. The ability to use HDR is dependent on whether or not the art assets needed to do the job are available.
In order to push this test to the max, the highest settings possible were tested. This includes enabling the reflect all and full HDR options. Our first test was run with no antialiasing and trilinear filtering, and our second set of numbers was generated after 4x antialiasing and 8x anisotropic filtering were enabled.
From the looks of our graph, we hit a pretty hard CPU barrier near 80 frames per second. The 7800 GTX sticks the closest to this limit for the longest, but still can't help but fall off when pushing two to three megapixel resolutions. The 7800 GT manages to out-perform the X1800 XT, but more interestingly, the X850 XT and X1800 XL put up nearly identical numbers in this test. The X1300 Pro and the X1600 XT parallel each other as well, with the X1600 XT giving almost the same numbers as the X1300 Pro one resolution higher. The 6600 GT (and thus the 6800 GT as well) out-performs the X1600 XT pretty well.
Antialiasing changes the performance trends a bit, but at lower resolutions, we still see our higher end parts reaching up towards that 80 fps CPU limit. The biggest difference that we see here is that the X1000 series are able to gain the lead in some cases as resolution increases with AA enabled. The X1800 XT takes the lead from the 7800 GTX at 1920x1200 and at 2048x1536. The 7800 GT manages to hold its lead over the X1800 XL, but neither is playable at the highest resolution with 4xAA/8xAF enabled. The 6600 GT and X1600 XT manage to stop being playable at anything over 1024x768 with AA enabled. On those parts, gamers are better off just increasing the resolution to 1280x960 with no AA.
The 7800 series parts try to pretend that they scale as well as the X1800 series at 1600x1200 and below when enabling AA. The X850 XT scales worse than the rest of the high end, but this time, the X1600 XT is able to show that it handles the move to AA better than either of the X8xx parts test and better than the 6600 GT as well. The 6800 GT takes the hit from antialiasing better than the X1600 XT though.
Doom 3 showed NVIDIA hardware leading at every step of the way in our previous tests. Does anything change when looking at numbers with and without AA enabled?
Rather than test Half-Life 2 once again, we decided to use a game that pushed the engine further. The reworked Day of Defeat: Source takes 4 levels from the original game and adds not only the same brush up given to Counter Strike: Source, but HDR capabilities as well. The new Source engine supports games with and without HDR. The ability to use HDR is dependent on whether or not the art assets needed to do the job are available.
In order to push this test to the max, the highest settings possible were tested. This includes enabling the reflect all and full HDR options. Our first test was run with no antialiasing and trilinear filtering, and our second set of numbers was generated after 4x antialiasing and 8x anisotropic filtering were enabled.
From the looks of our graph, we hit a pretty hard CPU barrier near 80 frames per second. The 7800 GTX sticks the closest to this limit for the longest, but still can't help but fall off when pushing two to three megapixel resolutions. The 7800 GT manages to out-perform the X1800 XT, but more interestingly, the X850 XT and X1800 XL put up nearly identical numbers in this test. The X1300 Pro and the X1600 XT parallel each other as well, with the X1600 XT giving almost the same numbers as the X1300 Pro one resolution higher. The 6600 GT (and thus the 6800 GT as well) out-performs the X1600 XT pretty well.
Antialiasing changes the performance trends a bit, but at lower resolutions, we still see our higher end parts reaching up towards that 80 fps CPU limit. The biggest difference that we see here is that the X1000 series are able to gain the lead in some cases as resolution increases with AA enabled. The X1800 XT takes the lead from the 7800 GTX at 1920x1200 and at 2048x1536. The 7800 GT manages to hold its lead over the X1800 XL, but neither is playable at the highest resolution with 4xAA/8xAF enabled. The 6600 GT and X1600 XT manage to stop being playable at anything over 1024x768 with AA enabled. On those parts, gamers are better off just increasing the resolution to 1280x960 with no AA.
The 7800 series parts try to pretend that they scale as well as the X1800 series at 1600x1200 and below when enabling AA. The X850 XT scales worse than the rest of the high end, but this time, the X1600 XT is able to show that it handles the move to AA better than either of the X8xx parts test and better than the 6600 GT as well. The 6800 GT takes the hit from antialiasing better than the X1600 XT though.
Doom 3 showed NVIDIA hardware leading at every step of the way in our previous tests. Does anything change when looking at numbers with and without AA enabled?
Facebook
Twitter
RSS
93 Comments
View All Comments
TheInvincibleMustard - Friday, October 7, 2005 - link
Hard|OCP didn't include the X1800XT in their review for precisely the reason you point out -- it's nowhere near available yet.-TIM
Xenoterranos - Saturday, October 8, 2005 - link
I think the prob was that they couldn't get one either. That's one of the advantages of being as big as Anandtech. Hard OCP is big mind you, just not ANANDABIG!DerekWilson - Friday, October 7, 2005 - link
By that logic they shouldn't have reveiwed any of the cards -- we can't buy any yet.Honestly, we aren't an advertising agency, we aren't trying to sell anyone anything. We are a hardware review site. We review new hardware and technology as we are able to get our hands on it. That does include guiding our readers towards making good purchasing decisions at times, and the lack of availability is definitely a factor in that recommendation.
But we can't be expected to ignore hardware we get just because it's not available. In fact, some of our best articles are on products that we got our hands on from 3rd party contacts months, weeks or days before the product is launched. For instanc, we got CrossFire parts a long time before ATI sent us product. The products wouldn't be available for months, but we wanted to review them anyway. The reason is because the hardware is interesting and people want to know the preformance characteristics and details of new products.
We certainly won't recommend the X1800 XT to anyone until we see it ship with the 625/1.5 clock speeds we tested at a reasonable price.
There are two sides to our philosophy on reviews, but our curiosity usually gets the better of us.
Also note that we didn't include the X1800 XT in all of our benchmarks in the launch article for the purpose of deemphasising it. Quite a few people complained about it being left out.
In any case, we want to bring readers the articles they want to read. And I'm not convinced that you guys really want us to leave out hardware. If we didn't test the hardware we had available, we would not only be avoiding telling the whole story, but we would be assuming our readers didn't have the sense to realize what is available for purchase and what is not.
You guys are smarter than that :-)
TheInvincibleMustard - Friday, October 7, 2005 - link
Oh true enough, and I'm not calling into question any of AT's journalistic integrity or anything like that. I do readily admit that I get as drool-y over The Next Big Thing(tm) as the next hardcore techno-enthusiast, but I also find it very frustrating (along with many others) that products get a "review" slapped on them even if they're not shipping ... for that matter, I seem to recall a review this past winter about a product that never even shipped.I'm fairly certain that I'm not alone in saying that I like the fact that AT (and other sites, like Hard|OCP, TechReport, etc.) put out numbers on something at the very first available opportunity (ie, when the NDA lifts). On the other hand, I'm also fairly certain that I'm not alone in the fact that it's very frustrating to us, as consumers, to see products being reviewed that won't be out for weeks, if not months (if ever). I'm also fairly certain that I'm not alone in believing that hardware manufacturers will continue to perfom these "paper launches" so long as major hardware review sites (such as AT) treat a non-shipping product in the way that they currently are.
In my personal opinion, I would absolutely love it if a major site, such as AT, would simply do away with reviews of products that consumers cannot buy. Purchase an X1800XL from NewEgg (or wherever) once it becomes available, and conduct a full review of it then. Leave the non-shipping versions as technology previews, or sneak peeks, or early looks, or whatever it is you'd like to call them. Video game review sites, such as Gamespot, make it a point to very much so distinguish between a "late-beta hands-on" or "release copy" and a "retail shipping product," and if they review a game and I like it, I can head on over to EBGames or wherever and purchase it that same day. That's the sort of approach that I feel would be most appropriate to curtail these paper launches ... so long as manufacturers think they can continue to get away with a full indepth comparison and review of a non-shipping product, they'll continue to do so. Personally, I think that ATi was not slammed hard enough for not having the parts available on store shelves as of earlier this week.
-TIM
bob661 - Saturday, October 8, 2005 - link
Speak for yourself bro...lol....I like knowing about the product BEFORE it ships because people ask me about these products and I can inform them WELL before the instant gratification kicks in. Feed me Seymour!!!!DerekWilson - Saturday, October 8, 2005 - link
How do we balance doing an architectural/technology article without performance tests of hardware?The problem is that a review looks like tech article or a preview on sites like ours ...
It's a bit easier to make the distinction with a game. Much can change even after release. But with hardware, ATI is locked into the silicon and everything but the clock speeds and memory sizes are fixed. If you clock an R520 part at 625/1.5 you will get the performance we showed. The only thing we would really like to do (that we can't yet because ATI doesn't build it into their drivers) is underclock the cards and look at the cheaper skus or do a frequency scaling article. That way we would end up covering something the end users will eventually see.
I agree that there is a problem in general with paper launches, but I think that the public has not been fooled this time. Especially after what happened last winter with the X800/X850 series and this past summer with CrossFire. On our end, its difficult to balance a solution.
TheInvincibleMustard - Saturday, October 8, 2005 - link
No, I do see your point of reference, and agree with your statement that it is "easier to make the distinction with a game." I guess that I'm just frustrated that what you folks are reviewing is not what I'll be buying (completely ignoring that whole "golden sample" mumbo-jumbo) right when new technology is launched. In fact, it was your review of the eVGA 7800GTX that helped me make my decision to purchase that card instead of a 6800GT that I was considering. It wasn't the review of the Reference Board that made me wish to purchase the 7800GTX, it was the review of an actual, shipping product, as that review drove home (to me, at least) much more so exactly what my money is going towards and what I'll be getting from spending that much on a piece of hardware.I've seen this extend beyond the graphics card segment, as well. Wesley's review of the Reference Board for the Crossfire ATi motherboards included high praise for what the chipset is capable of (which is well-deserved, aside from the USB performance, which I seem to recall ATi stating that they'd fix in this particular release) ... ATi actually went so far as to reference the article in their Q4 financial statement. While I realize that this is ATi's spin on things and something that is probably beyond the control of anyone at AT, it's frustrating to me to think about how people that aren't "in the know" see that quote coming from AT and think therefore that the Crossfire mobo is the Best Thing Since Sliced Bread(tm), when the actual review was of a Reference Board -- everyone has seen things in the past that are similar to this, such as SiS making very excellent chipsets that perform very well but end up getting delegated as the "value" segment for motherboards and therefore don't get the recognition that they deserve with retail shipping products. Is the Reference Board review incorrect? No, I don't think so, but I have a hard time acknowledging the fact that the Crossfire Reference Board is truly indicative of retail shipping products, which, on the surface at least (ie, without an indepth look into the review), it appears to be. Note that I am aware of the fact that DFI is, I believe, going to be taking this chipset and performing their voodoo magic on it, which will more than likely make it into a monster of a motherboard, but that's in the future and is not what Wesley's review contained.
Now, I certainly am also well aware of the fact that motherboards are a very different beast when it comes to manufacturing and the capabilities and performance bulit into it at that stage (eg, DFI vs Chaintech with an nForce4 chipset), but in my opinion the principle is the same -- what you're reviewing is not truly indicative of what the consumer will be purchasing, and for that, I'm saddened.
This, of course, in no way means that I'm going to be leaving AT to go read somewhere else. I've been a loyal reader of AT for probably five or six years now, and can remember reading about the 1.13GHz P3 fiasco, so from at least that timeframe. I certainly plan on continuing to read AT into the future, and have based many of my computer purchase decisions upon what I read in reviews at this site, and will probably continue to do so.
I guess, basically what it boils down to, is that I'm frustrated that it seems somehow acceptable -- to a certain extent -- that these "paper launches" are tolerated. I hope that my point of view is a bit more clear to you, now, and while I realize that my voice does not have a huge weight attached to it in the large scheme of things, since a forum is provided for feedback, that's what I'm doing.
On a much more positive note, I am extremely happy that AT is one of those places where there is a nice direct connection (these comments) between the site operators/writers and the site visitors. In my opinion, nothing harmful can come from improved communication, which is something that I truly enjoy about this site.
Keep up the good work, everyone at AT.
-TIM
BikeDude - Saturday, October 8, 2005 - link
I've been holding off upgrading my graphics system for several months now. This review helped me realise the next big thing is still a way off, unless nVidia rushes a 7800Ultra out in time for the holiday season. (I'm also going for that Apple monitor, hence I'd like a card that can deliver on resolutions higher than 2048x1536)Knowing what's down the road a month or two from now definitively helps making a buying decision today. Yes, the schedule might slip, but someone buying a 6800Ultra in April would probably have liked to know about the 7800GTX' pending launch.
Having access to more information can never be considered bad. The problem here is that the manufacturers enjoy giving us good news ("Next Generation part performs well!"), but seldom tells us the bad ("Next generation part delayed due to unforseen problem"). But IMO I prefer that to having no news at all!
JarredWalton - Saturday, October 8, 2005 - link
So, correct me if I'm wrong, but your basic issue is with the reference product articles being called "reviews". If we change the title to "preview" then you're happy? At least from my perspective, it really doesn't change anything. I mean, a review of the technology and potential performance or a "preview" of the same thing... "A rose by any other name would smell as sweet," right?Anyway, I don't really have any issues with changing articles like this to have "preview" in the title, but that's not my call. Take it up with the big man (Anand) if you'd like. :) I just assume that everyone realizes that when we say the product is not yet available in stores, then we're not giving a recommendation to buy any specific hardware until it shows up.
TheInvincibleMustard - Saturday, October 8, 2005 - link
That is it, to a certain extent, but going further than that allows the change for whether this new part is going to be as-heavily-focused-on as a retail shipping product? Forgive me for beating a dead horse, but if we go back to the game (p)review concept -- nearly all big-budget games would get a hands-on late-beta or release-candidtate preview, as would be expected from the readership (they want to know what's going to be going on), and AT should be no different with hardware. However, the depth of coverage from the gaming site for the preview is nowhere near the amount that they put into the retail shipping product -- previews of big-budget games could be three or four paragraphs and a few more screenshots, maybe some beta gameply video, while the final review would be several pages, some form of ratings or scoring system for direct comparison with other products, usually an indepth video review featuring gameplay elements, and the like. In other words, there's a very clear distinction between what constitutes a preview and what constitutes a review, not only from the title of the article but to how much and what type of information is actually packed into the article.Does this mean that I think AT or other sites should not publish any information if there's no retail shipping product? No, of course not, information is helpful. However, when the line between a "technology preview" and a "retail product review" becomes blurred -- not only in terms of the title of the article but in the content of the article as well -- then that is when things seem to be not so "rosy" (pun intended).
I hope that this post has made my point a little more clear, and if not, I'll try again. Just this thread, though, is exactly my point about what makes AT great -- the people that are doing the stuff on the site are willing to listen to me, one of the million lowly peons that click through every day. If everything else about the site changes, I most sincerely hope that at least this aspect of it would remain the same.
Thanks again for your time.
-TIM