Is the 6800NU *really* faster than the X800XT in Doom3?

[SickBeast]

Some people on here have been shamelessly posting that the 6800NU beats out the X800XT in Doom3 (*cough* Rollo *cough*). I have been reading the new AnandTech article on Doom3 and the 6800NU didn't beat the X800XT in a single test.

Am I missing something? Are there benchmarks floating around which suggest that the 6800NU is faster?

I realize that the 6800NU put up an excellent showing against a graphics card which costs $200 more than it, but I don't understand why people are suggesting that it is faster than ATi's top card.

Enlightenment would be nice. If you're going to make a claim, make sure you have some benchmarks to back it up.

 

[Corporate Thug]

i trust anandtechs benchies over anyone else, so if they show the x800xt beating the 6800, then i dont doubt it

 

[rbV5]

Anandtech used a single AMD platfoerm to bench the cards, where HardOCP used a variety of platforms in their buying guide.

On the Northwood platform, the performance is very close between the 2 cards @ 16x12 NoAA 8XAF ( 6800 X800XT

 

[ss284]

Looking at the vr-zone benchmarks, which I believe to be the least biased, shows the 6800 tying or slightly slower than the x800xt with no filtering enabled. Once AA/AF is enabled though, the 6800 falls inbetween the x800pro and the x800xt.

Its still pretty disheartening to see such a huge variance in bechmarks.

-Steve

 

[Ackmed]

Not when you turn up the res and detail its not. For the money though, its probably the best Doom3 card of these new cards.

shady06, I used to trust Anand a lot more than I do now. With the past few mistakes, trust starts to dwindle. This review sends up a few flags, not that Im saying they falsified the results, but I am just confused on a few of them.

Question 1

How does the 9800 Pro 128meg get 5.5 more frames than the 9800XT? And how does the 9800XT get the exact same frames as the 9700 Pro? Does that seem off to anyone else?

Question 2

In High Quality and High Res the X800XT/PE gets five more frames than in High Quality and Medium Res? The settings seem to be the same to me, with the only difference that High Res is 1600x1200, and the Medium Res is 1024x768.

If there is some logical answer for this, Im all ears. I dont claim to know everything, but just by looking at the graphs, it seems like they are off. I just cant understand why the 9800 Pro would beat the 9800XT, and why the 9800XT would score the same as the 9700 Pro. I also dont understand why the XT/PE would get a higher score, with a higher res. Anyone?

 

[DAPUNISHER]

Hey Ackmed,

Derek was up all night and probably a little punchy so I don't doubt he could have made a mistake or 2. What seperates AT from many sites is that they will find and fix any errors pointed out, take full responsibility for them and issue the appropriate retractions if need be. They are also always open to our input. Consequently I still implictly trust this site for hardware data, I just don't repeat it as fact until I'm certain they have already went over any questionable content with the proverbial fine tooth comb

 

[ss284]

Good point Ackmed, but I think these anomalies are more attributable to the test setup, and how the benchmarks were carried out. That said, I also dont trust anandtech 100% anymore. Some of the smaller enthusiast sites, especially those in asia, I feel, are less biased.

-Steve

 

[Ackmed]

Im not calling him bias, or implying that they results are fake. I am just confused as to how these results can come up like this. There have been a lot of errors in the video card reviews lately, at least to me.

 

[vshah]

ackmed, i believe the 9800xt/pro 128 thing is a typo, the graph makes a lot more sense if those two labels are switched around.

-Vivan

 

[DAPUNISHER]

Originally posted by: ss284
Good point Ackmed, but I think these anomalies are more attributable to the test setup, and how the benchmarks were carried out. That said, I also dont trust anandtech 100% anymore. Some of the smaller enthusiast sites, especially those in asia, I feel, are less biased.

-Steve

You and I are diametrically oppossed on this one. I find the testing methodology, as well as the technical background of many of the smaller sites reviewers very poor compared to the top five or 6 sites.

 

[KevinH]

Originally posted by: DAPUNISHER

Originally posted by: ss284
Good point Ackmed, but I think these anomalies are more attributable to the test setup, and how the benchmarks were carried out. That said, I also dont trust anandtech 100% anymore. Some of the smaller enthusiast sites, especially those in asia, I feel, are less biased.

-Steve

You and I are diametrically oppossed on this one. I find the testing methodology, as well as the technical background of many of the smaller sites reviewers very poor compared to the top five or 6 sites.

By the same token, inaccuracies =/ bias. Bias implies they are in somebody's pocketbook or favor one company over another for wahtever reason. Look at how AT has historically picked the better contender at the time. Bias is WAY overused.

 

[GeneralGrievous]

They are close, but the XT is faster. I think RVBS provided the best comparison, though that was an overclocked 6800. 4.9 catalysts might help.

 

[RussianSensation]

Doom 3 Performance
Hopefully this article will help to shed more light on the matter.

I think that even if 6800nu is at least as fast or faster than X800xt in just 1 game, that is not an indication of its overall performance. In fact, in most games it is only as fast as a 9800xt and often slower due to its low subsystem memory.

Recently I read the most comprehesive videocard comparison so far which proved to be very very helpful.
The Fastest Graphics Cards of Summer 2004 - Testing with 35 Games (!)

This article shows that X800xt is often either as fast or slightly faster than 6800ultra, 6800GT is an overall better card than X800pro, and that 9800Pro and 9800xt are better choices than 5900xt and 5950ultra respectively.

The article goes on to conclude that GT is a very favourable card because it costs $100+ less than the top end cards and can overclock to Ultra speeds. HL2 performance was also rather suprising considering that Nvidia (geforce 6) cards did much better in pure performance modes. Also my highly anticipated game, STALKER, showed that either top end card will handle the highest level of quality at more than enough frame rate levels.

And finally, there are different types of people who tend to take 1 small fact, occurance, and so on, and exaggerate the importance of that event, yet ignoring the overall picture.

For instance, saying that PS3.0 is important now is contradictary in nature as X800xt beats any Nvidia card (even with SM3.0 enabled), while lacking PS3.0 support in the only game that supports such a feature. Also telling people to not buy a videocard based on the performance of 1 game alone is at best questionable. We have to consider that by the time the next game that utilizes Doom 3 engine comes out, most people who spend $400 US on a videocard, will acquire the next generation (NV50/R520 card, etc.) card. The best card to a person is the one that plays his/her games the fastest I suppose. Other features such as dual-dvi support, low psu requirement, 1 or 2 slot design, quiet fan, stability of drivers and availability of the card should be considered.

If I was buying a videocard now, I would never spend $300 on 6800nu because for $100 more you can get a card (GT) that is often 2 times faster with higher quality settings. I'd either get a 9800Pro or 6800GT, but nothing in between.

 

[gururu]

Originally posted by: Ackmed
Not when you turn up the res and detail its not. For the money though, its probably the best Doom3 card of these new cards.

shady06, I used to trust Anand a lot more than I do now. With the past few mistakes, trust starts to dwindle. This review sends up a few flags, not that Im saying they falsified the results, but I am just confused on a few of them.

Question 1

How does the 9800 Pro 128meg get 5.5 more frames than the 9800XT? And how does the 9800XT get the exact same frames as the 9700 Pro? Does that seem off to anyone else?

Question 2

In High Quality and High Res the X800XT/PE gets five more frames than in High Quality and Medium Res? The settings seem to be the same to me, with the only difference that High Res is 1600x1200, and the Medium Res is 1024x768.

If there is some logical answer for this, Im all ears. I dont claim to know everything, but just by looking at the graphs, it seems like they are off. I just cant understand why the 9800 Pro would beat the 9800XT, and why the 9800XT would score the same as the 9700 Pro. I also dont understand why the XT/PE would get a higher score, with a higher res. Anyone?

on behalf of the 9800xt.9700pro comparison. its very close at those settings. at 1280, my 9700pro gets close to or higher than what some users are reporting with 9800pros. at 1024, i get 35 fps and others report 41 fps with 9800pro. I think, the two cards are very very close in this game, and this might be a driver issue.

 

[Illissius]

In the review at xbitlabs, the 6800 is usually faster than the X800XTPE w/o AA/AF, and is usually slower with them (but there are exceptions to either case).
(Not saying I trust these results, or that I do not, or anything at all, just that this is one of the places where people are probably getting the impression that the 6800 is faster.)

 

[Pete]

Originally posted by: DAPUNISHER
What seperates AT from many sites is that they will find and fix any errors pointed out, take full responsibility for them and issue the appropriate retractions if need be.

Erm, yes, AT has certainly improved its graphics card coverage with the addition of Derek as the dedicated 3D card reviewer, but AT has certainly had its problem(s) WRT article errors and corrections. No site is perfect, especially in such a quickly-changing field.

 

[rbV5]

but I don't understand why people are suggesting that it is faster than ATi's top card.

Probably for the same reason "others" were suggesting that the 6800 owners should be "sad" they would be stuck at medium detail settings....

 

[GeneralGrievous]

Probably for the same reason "others" were suggesting that the 6800 owners should be "sad" they would be stuck at medium detail settings..

Well others suggested that because some id rep said 256 mb of ram was required for High Quality.

 

[rbV5]

Originally posted by: GeneralGrievous

Probably for the same reason "others" were suggesting that the 6800 owners should be "sad" they would be stuck at medium detail settings..

Well others suggested that because some id rep said 256 mb of ram was required for High Quality.

Exactly, and I'm guessing Rollo didn't try D3 on an X800Xt either.

 

[nRollo]

Originally posted by: SickBeast
Some people on here have been shamelessly posting that the 6800NU beats out the X800XT in Doom3 (*cough* Rollo *cough*). I have been reading the new AnandTech article on Doom3 and the 6800NU didn't beat the X800XT in a single test.

Am I missing something? Are there benchmarks floating around which suggest that the 6800NU is faster?

I realize that the 6800NU put up an excellent showing against a graphics card which costs $200 more than it, but I don't understand why people are suggesting that it is faster than ATi's top card.

Enlightenment would be nice. If you're going to make a claim, make sure you have some benchmarks to back it up.

I don't think I've ever said the 6800NU is faster than a X800XT PE at Doom3, only as fast. Which, at some settings, it is. Obviously 16X12 4x8X isn't going to be one of them.

Oh well- it's time!

My XFX 6800GT arrived today, time for us to all learn the difference between the 6800NU benches I posted and the 6800GT.

 

[SickBeast]

Originally posted by: Rollo

Originally posted by: SickBeast
Some people on here have been shamelessly posting that the 6800NU beats out the X800XT in Doom3 (*cough* Rollo *cough*). I have been reading the new AnandTech article on Doom3 and the 6800NU didn't beat the X800XT in a single test.

Am I missing something? Are there benchmarks floating around which suggest that the 6800NU is faster?

I realize that the 6800NU put up an excellent showing against a graphics card which costs $200 more than it, but I don't understand why people are suggesting that it is faster than ATi's top card.

Enlightenment would be nice. If you're going to make a claim, make sure you have some benchmarks to back it up.

I don't think I've ever said the 6800NU is faster than a X800XT PE at Doom3, only as fast. Which, at some settings, it is. Obviously 16X12 4x8X isn't going to be one of them.

Oh well- it's time!

My XFX 6800GT arrived today, time for us to all learn the difference between the 6800NU benches I posted and the 6800GT.

:jealous;

 

[DAPUNISHER]

Originally posted by: Pete

Originally posted by: DAPUNISHER
What seperates AT from many sites is that they will find and fix any errors pointed out, take full responsibility for them and issue the appropriate retractions if need be.

Erm, yes, AT has certainly improved its graphics card coverage with the addition of Derek as the dedicated 3D card reviewer, but AT has certainly had its problem(s) WRT article errors and corrections. No site is perfect, especially in such a quickly-changing field.

I certainly didn't intend to imply otherwise. They do a better job than other high traffic sites such as Tom's and [ H ] IMO, and that makes them elite in my book.

 

[Gamingphreek]

Hey guys give em a break everyone makes mistakes. Everyone has made a mistake at some time. Lay off on the insults a bit.

Anyways... the NU is faster than the 9800XT in all cases. The 256mb of memory is not needed because neither card is fast enough to set the options that do need it.

Yes, in all the D3 performance reviews i have read have the 6800NU tieing with the X800XT and slightly slower in most cases. However the X800Pro doesn't come close (in most cases) to the NU. The GT and the Ultra (and the UE if you will), lead the pack in everything there.

Also i would like to know how some people know that those orders swap when AA and AF are enabled. yes i know Nvidia is not good at AF but the cards dont swap ecause 8x AF is always on in High Quality IIRC. Now more specifically it was stated this occurs when AA and AF are enabled. How in the world would anyone know that. The highest settings that are playable on any machine are 16x12 Ultra Detail, 2xAA. And that is not with the ATI cards and that is with a A64FX-53 with 2GIG!!! of memory. SO how would the cards swap when none of them are fast enough to max out everything.

-Kevin

 

[GeneralGrievous]

Well, anand updated his tests with the 4.9 catalysts. The XT is now ahead of the 6800 nu by a decent margin, even moreso at ultra quality.

 

[DAPUNISHER]

Originally posted by: GeneralGrievous
Well, anand updated his tests with the 4.9 catalysts. The XT is now ahead of the 6800 nu by a decent margin, even moreso at ultra quality.

I wouldn't call it decent given the price difference between them