Radeon X800PRO will beat NV40 Ultra

[Druga Runda]

well do I need to say more

It's from le Inq

http://www.theinquirer.net/?article=15406

fairly new so I hope it wasn't posted yet anywhere else.

WISE BIRDS tell us here in Vienna that ATI and at least some of their knowledgeable partners don't feel so bad about NV40 Ultra. Even though the NV 40 Ultra, Geforce 6800 Ultra is a very fast card that outperforms everything else on planet for the time being, the ATI next generation chip might end up even faster.

The canaries are singing that 12 pipelines, 475MHz/950MHz card with 96 bit precision and PS 2.0 shader only will end up faster then 16 pipelines, 400MHz /1100MHz cards with 128 bit precision and PS 3.0 shader model.

ATI knew for some time it would lose out feature wise, but stays committed to bring the faster technology, without some of the extra features.

The R400 card - now renamed to R500 - has support for all these nice features but got postponed but that is the chip with 128 bit precision and PS 3.0 support from ATI. Until that chip dawns, later this year at the earliest, ATI stays PS 2.0 and 96 bit precise only.

But Nvidia seems to be keeping some of its powder dry too. We hear about some faster NV40 chips that Nvidia is saving for later. µ

 

[Tim Murray]

Why is Fudo using 96-bit and 128-bit precision instead of FP24 and FP32? Oh Fudo, I slap you with a trout labelled "Clue Stick." Only marketing people say 96 and 128-bit precision. Everyone else says FP24 or FP32.

 

[Kombatant]

It's as i've said in the past... whatever happens, the Inq will be right in any case... it has posted all possible combinations of facts

 

[Tim Murray]

It's as i've said in the past... whatever happens, the Inq will be right in any case... it has posted all possible combinations of facts

So true. My all-time favorite Inq comment was where Fudo said that the NV40 was capable of 32 color-without-Z pixels per clock. Fudo, you're such a silly goose.

 

[Psikotiko]

Ohhhh!! The usual monday morning Inquirer "story".

Next monday, the parhelia 8x will beat X800 and Nv40.

 

[Druga Runda]

It's as i've said in the past... whatever happens, the Inq will be right in any case... it has posted all possible combinations of facts

But this is too much... Pro faster than the current Ultra?

I'd like to see that happen... and than a clock speedup on both ends.

But just given the fact that the XT comes out in a month should be a hint enough that PRO is too slow for NV40.

 

[mikechai]

Don't you find it amusing that every new INQ post, there's a thread for it ...

 

[indio]

Don't you find it amusing that every new INQ post, there's a thread for it ...

I find it funny for every new thread at B3D there's an INQ post for it.

 

[mikechai]

Don't you find it amusing that every new INQ post, there's a thread for it ...

I find it funny for every new thread at B3D there's an INQ post for it.

Owned.

 

[jimmyjames123]

I'd say it is more realistic to say that performance lead will depend on what game is tested, what AA/AF level is used, and what resolution is used.

 

[Druga Runda]

I'd say it is more realistic to say that performance lead will depend on what game is tested, what AA/AF level is used, and what resolution is used.

One way or another I am still suprised as theoretically the PRO should be inferior to NV 40 as a chip and the memory on board.

So - well if it is faster in some games on high resolutions esp with AA/AF

I'd expect it to be competitive in CPU bound situations, as usual, and that will be plenty as both boards are very good comparing to the current crop of games.

But where it really matters NV40 should pull ahead 10-15+ % , if the architectures are of similar efficiency, now add or deduct depending for which architechture the game is better suited, but should be similar as both boards are geared towards similar functionality unlike R300/350 vs NV30/35

 

[Diplo]

...the ATI next generation chip might end up even faster.

What a total non-piece of journalism. Of course it might be faster, but then it might not. How can the Inquirer ever be wrong with such amazing insights as that?

 

[991060]

I found in traditional games, 6800U is severely bandwidth limited. So X800Pro could beat 6800U provided it's equipped with very faste memory, say 700-800MHz.

 

[Druga Runda]

...the ATI next generation chip might end up even faster.

What a total non-piece of journalism. Of course it might be faster, but then it might not. How can the Inquirer ever be wrong with such amazing insights as that?

well I guess the title of the article clear that out - read: will be faster - at least this is what Fudo thinks for unknown reasons

 

[Druga Runda]

I found in traditional games, 6800U is severely bandwidth limited. So X800Pro could beat 6800U provided it's equipped with very faste memory, say 700-800MHz.

but isn't the current unofficial spec 450 mhz... that is 900DDR, as opposed to 550 'official' for NV40.

but you never know if faster mem is true, it would be the PRO's advantage.

But if the PRO is competitive to 6800U where does it leave the plain 6800 on 128bit bus? Aren't they the same price range?

 

[Tim]

...the ATI next generation chip might end up even faster.

What a total non-piece of journalism. Of course it might be faster, but then it might not. How can the Inquirer ever be wrong with such amazing insights as that?

There is a huge difference between the writers at the Inq, Fudo rarely post anything of value, neither does Eva Glass. Charlie on the other hand often posts some good stuff and he right mush more often than he is wrong.

The Inq needs a much more critical editor because they actually post some good stuff pretty often, but it easy to miss it because of all the crap.

 

[Psikotiko]

But if the PRO is competitive to 6800U where does it leave the plain 6800 on 128bit bus? Aren't they the same price range?

128 bit bus???
6800 non ultra has a 256 bit bus.

 

[991060]

but isn't the current unofficial spec 450 mhz... that is 900DDR, as opposed to 550 'official' for NV40.

but you never know if faster mem is true, it would be the PRO's advantage.

True, I just feel ATi won't put that fast memory on X800Pro coz there's X800XT. If X800Pro is already using the fastest memory, how can you expect X800XT to be faster enough into order to justify the additional 100$?

 

[Druga Runda]

IMHO withouth them, or this kind of 'rumors are here to be reported' attitude Inq would be just another boring tech mag. It's because of this no compromise attitude that I read them, and well they might be wrong and they are fed FUD sometimes by the people on the inside, but I am pretty certain that they all only post when they believe that the stuff has credibility.

And that is what keeps the rumor mill turning. I am sure without Inq 1/2 rumors would be confined to much smaller audience and make the whole waiting fopr new products period much more boring...

 

[DemoCoder]

Either the the 3dMark03/ShaderMark figures I have seen are horribly wrong, or the 12-pipe pro won't be able to top the NV40 but will lag about 15% behind. I think everyone expects the 16-pipe R420 to win, atleast until Nvidia goes through a clock-rev.

I'm actually hoping it (R420) will win. I've already made the decision (last year) that I'm not upgrading my system until BTX PCI-E mainboards become available. In which case, I will buy a PCI-E card. That means, I will be buying either the R423 or an NV4x with native PCI-E. I expect the NV40 with native PCI-E will be tweaked for higher yields and clocks too. If not, then hopefully a 500Mhz+ R420 will scare NVidia into action.