Excellent work Sander, and you're vindicated indeed. My hat's off to you, you stayed calm and reasonable through all of this and haven't slanted ATI in any of the subsequent articles. I now put more value into my membership here as I know you can be trusted to speak the truth, or your mind, about things that matter to your readers. Keep up the good work!
Want to enjoy fewer advertisements and more features? Click here to become a Hardware Analysis registered user.
jason, if you need any help with anything gimme a call, if i dont know i can probably refer you to someone that knows about your subject. Also what it appears is ATI has looked to their past for improvements (before the x800 series) when the R300s and R350s/360s ruled the roost.
Its very interesting to see how the various review sites have reacted to this launch. Some have been bitterly disgruntled (and rightly so) about ATi' paper launch of CrossFire and now of over half of their new product line. Some have had nary a whisper about this major faux pas. Now, after reading most of the major sites' reviews of the new technology, it feels to me that biases are coming out more strongly than ever before. My initial feeling is that at least for the high midrange, such as the 7800GT vs X1800XL, the cards are fairly evenly matched. I expect this to change, however, with Nvidia's Rel80 driver. Most hardware sites are spelling this out pretty clearly, with varying levels of harsh words for ATi's questionable launch tactics. However, some, like Hexus.net, seem to be falling over themselves to praise ATi, proudly proclaiming the non-existant X1800XT as the 'fastest card on the planet' and largely excusing the fact that it doesn't exist outside the reviewer's own hands. This has been a very tumultuous time for the hardware review community at large, with things getting much more 'personal' than ever before. It'll be interesting to see where it goes.
I think to be fully vindicated the benchmarks have to be shown to match those of the current batch of reviews, though the article implies they do, I really don't think they do. Look at spliter cell for example. The card went from being beaten by the 7800GTX to beating it by nearly 25%. Also the scaling with resolution changes does not match the behaviour of the x1800xt in the new benchmarks. What the hardware analysis benchmarks do show is a card that tracks an x850xt very closely. If I was asked I would guess it was an overclocked x850xt too.
You need to factor in that the GeForce cards used the 71.72 driver, that's ancient by today's standards, the rel 80s offer a significant (10 to 20%) increase in performance. All the reviews that you've read today use this driver. What you should be looking for is trends and they're all there, for example look at the TechReport review, Scott uses almost the same demos and the same script to benchmark these cards.
Sander, the Tech Report benchmarks were the first I looked at and I was looking at the trends. The Hardware Analysis benchmarks show the x1800xt trailing splinter cell by about 5% at 1024x768 and getting progressively further behind as the resolution increases. If as you say the new nVidia driver gives the card a 10-20% boost in performance why is the x1800xt now beating the 7800gtx by 20% without the drop off in performance as the resolution increases?
I don't have the answer to that question, and I know that we used a custom demo on the Splintercell benchmarks, so you cannot compare directly to Scott's. However it is clear that the results we've shown are in line what the actual reviews are showing. Also keep in mind that ATI upped the clocks by 25MHz on the core and 100MHz on the memory and is several builds further with their driver since we posted ours.
Sorry, but I don't see a correlation at all. Another example farcry. The drop off of 7800gtx, x850xt and x1800xt are pretty much linear in your benchmarks. On other benchmarks the x1800xt virtually no reduction of framerate as resolution increases. This would seen to show that the frame rate limit of your x1800xt review is a fillrate limit where and the other reviews are hitting a vertex throughput limit. No noticable fillrate limit even with 4xAA and 16xAF. I am pretty up on 3D tech, been involved now for 15 years, I won't say where on here but I assume you can check my email address Sander?
Ah, yes. Another jewel of mainstream, objective, professional reporting.
It would seem, in fact, that yes. An X1800 XL performs right next to a 7800 GT, most ironically in the benchmarks you chose to pit it in. It is neck-and-neck with the Nvidia architecture in Doom 3, Half-Life 2 and Splinter Cell: Chaos theory.
I just spend most of my day writing 3D engines, analysing and improving performance of 3D games on PCs and consoles. On the software side that is, not hardware. ATI, nVidia I don't care, the only system I find hideously ugly is the PS2
I agree with what Darren Bremner said today Wed 5 Oct 2005 at 11:34:
« to be fully vindicated the benchmarks have to be shown to match those of the current batch of reviews...
...If I was asked I would guess it was an overclocked x850xt too. ».
Yes, so far the benchmarks published by Sander have no proofs, and could be completely forged on paper by extrapolating the scores of an x850xt.
Moreover, the initial quibble left me doubting because Sander Sassen published truncated email exchanges, omitting some messages that must have been important according to his own account (see Summary, http://www.hardwareanalysis.com/content/article/1814/ ), while presenting the remaining ones with some bias.
« On August 15th I received an email from Andrzej Bania asking whether I was interested in an interview with a mainboard ASIC engineer. I declined the offer due to lack of interest amongst our readers for such an interview and asked for more information about R520 instead. Jokes were made back and forth. »
Which doesn't clear ATI from their own reactions on forums, where ATI hid still more than Sander, while unjustly inflating the quibble.
Too bad, Sander often has interesting views and informations - and sure ATI is a great company too.
Back and forth, back and forth. The bickering is like playing pong - from one side to another, nothing changes. Sander has numerous points (which I believe he has proven) that stand his ground fairly well. Mr. 3D-extraordinaire, on the other hand, also seems very well spoken and informed. What most people seem to miss is the big picture; ATI and nVidia are corporations out for one thing - to make a profit. Some will argue that these companies have other goals. Surely you are correct, but they are driven by profit-seeking. That means that they must capture the attention of the public, their consumers, and retain a hold on them. Apple, for example, claimed that the G5 would outperforme most any other system on the market. However, while the benchmark scores themselves were not forged, they were not accurate displays of performance. Apple used some Mac-specific software to test the computers. Does that mean they lied? Not technically. But did they tell us everything they knew? Hardly. I believe the same can go for nVidia and ATI.
Way to go dude, ATi's "He's talking a load ofs**t, and look, he's a rapist child molesting murdering lucitherian monster that pleasures himself in a tent" slander tactics (though i doubt they went that far...) didn't stand up for long.
"The X1800 XL performed at a similar, albeit somewhat lower performance level versus a GeForce 7800 GT, but even a GeForce 6800 Ultra was faster in many circumstances. The Radeon X1800 XT, however, if we disregard the multi-GPU configurations, traded the top spot with a GeForce 7800 GTX in some tests (3DMark, Splinter Cell, FarCry, HL2) with about a 60/40 split in favor of NVIDIA. And in typical fashion the Radeons shine better with AA and Aniso Filtering enabled, but the X1800 XT's larger 512MB frame buffer certainly helped it in this area. In general, ATI's new cards performed better in Direct3D applications than they did in OpenGL applications, which has historically been the case for ATI's products. Overall though, we'd consider the GeForce 7800 GTX the "faster" all-around card in terms of general gaming performance."
So, whether the card succeeds will depend on one thing - how much they decide to sell it for. Much more than the 7800GTX and they don't stand much chance, and given it's a 90nm fabrication, they're going to need high yields to get the price down. At the moment the prices appear similar, so it looks like Nvidia still have a slight edge.
LOL Jason I love it.,
I have been reading your articles as they have been published and I am very pleased, hell they have even made me chuckle at the utter "F**K you factor to ATI", both for trying to cut you out of the game and you coming back with the first preview...
I know that hardware analysis is the best most truthful review site and aint buying into no marketing bullshit payoffs from the big names.
These articles alone have lead me to point many people to the website.
Keep up the good work, as I am loving this too much.
This site is the best! Personally, I think all naysayers should be banned, but Sander allows the freedom of speech. That right there should allow everyone to see that he just posts the facts as he sees them and lets everyone else do the same. If he was truly biased, (like *cough* driverhell *cough*), then he would be deleted all possible threads that didn't direct the light in a way that he feels it should be directed. That's one of the best parts about this site.
I don't think I understood you as well as I should (unclear?), so despite that I will make my statement anyways: The X1800XT is starting at $700 USD, meaning the prices, as they stand, are not "close". Plus, it will not be released until November the 5th. As my high-school friends would say, they are boned.