ATI X800XT PE catches up with Nvidia SLI score

[Creig]

Originally posted by: JackBurton
[ATI won't have the option for SLI on their X800 cards.

Alienware has developed a motherboard that will allow ANY two identical PCI-E video cards to run in parallel. In fact, I believe they mentioned a possibility of utilizing up to four cards.

 

[bunnyfubbles]

Originally posted by: gururu

Originally posted by: thelanx
There's an overclocked ultra that's not that far behind, if two of those were in SLI, it'd blow the XT score away. Then again, that's two cards vs. one, albeit there are overhead losses, still not exactly a fair fight.

except that you can't SLI two overclocked cards. they require perfect sync, and any discrepancy in performance caused by out-of-spec parameters would cause them, well, to be out-of-sync.


edit: actually I take that back, these guys say it is possible:
Pantherproducts

I was going to ask where you got that idea when all the 2x card solutions (nVidia's SLI and Alienware's array) don't require perfectly synced cards.

With Alienware you can plug in any two cards, and with nVidia's solution they balance the load to make sure each GPU is equally stressed (ie one renders more screen where there is less action going on while the other renders less screen with more action) Not to mention you should probably be able to equally overclock the 2 parts and then run them in SLI

 

[clicknext]

Isn't this thing done by an individual and not ATI? Why are people saying that ATI is bloating their score with overclocking?

 

[bunnyfubbles]

I think the idea is that ATI can now claim superiority because the record was obtained with their hardware. However people shouldn't get worked up over it either way, 3DMark shouldn't be considered as they solution for determining how well your system games while at the same time 3DMark isn't completely worthless, its still a very good application to hlep burn in your video card and test stability

 

[gururu]

Originally posted by: bunnyfubbles
I think the idea is that ATI can now claim superiority because the record was obtained with their hardware. However people shouldn't get worked up over it either way, 3DMark shouldn't be considered as they solution for determining how well your system games while at the same time 3DMark isn't completely worthless, its still a very good application to hlep burn in your video card and test stability

the one thing I think about 3dmark05, in its favor, is that it seems to be less dependent on drivers than some gaming benchmarks, such as D3. I still think Farcry is the best and most telling bench available. Even HL2 to me, seems a little biased towards ATI through drivers and features.

 

[futuristicmonkey]

Originally posted by: gururu
Even HL2 to me, seems a little biased towards ATI through drivers and features.

Even after all the time Valve used to optimize the garbage out of HL2 just so it'd run half-decently on NV3x?

 

[gururu]

Originally posted by: futuristicmonkey

Originally posted by: gururu
Even HL2 to me, seems a little biased towards ATI through drivers and features.

Even after all the time Valve used to optimize the garbage out of HL2 just so it'd run half-decently on NV3x?

well, what they basically did was to force it to run in dx 8.0

 

[Gamingphreek]

Well guys with 2 of the fastest card out there in a system dont you think we are a bit CPU limited?

ATI's chips are not nearly as fast as 2 highend cards, all it is is CPU limitations and also unequal playing field (different HW).

-Kevin

 

[ronnn]

Originally posted by: Gamingphreek
Well guys with 2 of the fastest card out there in a system dont you think we are a bit CPU limited?

ATI's chips are not nearly as fast as 2 highend cards, all it is is CPU limitations and also unequal playing field (different HW).

-Kevin

So how do you know how efficient sli is? I agree one would assume greater results than this, but maybe there are still major bugs to work outl

 

[Marsumane]

The logic behnd this is funny. Its like saying "my 2500+ AXP performs better then your FX53 in Doom 3 so u should buy the 2500+ if the two were the same price even though i oced it to a 520fsb w/ liquid nitrogen..."

I can see where nv has somewhat of an argument. If you will buy the ultra already, u may as well be abpt buy two of them cuz you are already spending alot of money on the setup. But overclocking is universal to all cards. It shouldnt be compared unless both companies offerings are equal. ATI's offering cant be made equal to the nv offering, so its more like nv has an option for the rich to get high scores. Its an option that isnt comparable to anything that ati has to offer. They maintain the lead through their offering (im assuming if they were oced they would), but still its not comparable to ati's, but instead just an option that is placed on the table that is above and beyond what most of us can afford. (sorry for this rant, i know its kinda repetative and monotonous 😛 )

 

[Gamingphreek]

Originally posted by: ronnn

Originally posted by: Gamingphreek
Well guys with 2 of the fastest card out there in a system dont you think we are a bit CPU limited?

ATI's chips are not nearly as fast as 2 highend cards, all it is is CPU limitations and also unequal playing field (different HW).

-Kevin

So how do you know how efficient sli is? I agree one would assume greater results than this, but maybe there are still major bugs to work outl

I already said in my first post that i think there are also bugs that need to be worked out. But the fact is it is CPU limited. 1 can be CPU limited in a lot of cases... 2 is most definately.

-Kevin

 

[ZobarStyl]

Why is everyone getting so worked up over 3DMark05 scores anyway? We all know the true test of how well a card performs is how many BungholioMarks it gets. :laugh:

 

[lordtyranus]

Originally posted by: Gamingphreek

Originally posted by: ronnn

Originally posted by: Gamingphreek
Well guys with 2 of the fastest card out there in a system dont you think we are a bit CPU limited?

ATI's chips are not nearly as fast as 2 highend cards, all it is is CPU limitations and also unequal playing field (different HW).

-Kevin

So how do you know how efficient sli is? I agree one would assume greater results than this, but maybe there are still major bugs to work outl

I already said in my first post that i think there are also bugs that need to be worked out. But the fact is it is CPU limited. 1 can be CPU limited in a lot of cases... 2 is most definately.

-Kevin

You're not CPU limited if the XTPE is scoring higher. It's clearly a combination of already poor 6800 3dmark05 performance with some bug in SLI.

 

[Drayvn]

Well the thing is to me....

Is the SLI config gives 7250 right

X800XT-PE gives 5750

So the extra $400 or more gives u 1500 points, to me thats just a waste of money...

But if we compare that to the 6800U thats 4750

So ur getting 2500 which is better. But still, its a bit of a waste of money.

This i think helps the argument that SLI might have a few bugs in it... maybe, dunno...

 

[DAPUNISHER]

Originally posted by: Drayvn
This i think helps the argument that SLI might have a few bugs in it... maybe, dunno...

I think that is the likely answer, it's just not ready for primetime tech yet.

 

[Childs]

Originally posted by: lordtyranus

Originally posted by: Gamingphreek

Originally posted by: ronnn

Originally posted by: Gamingphreek
Well guys with 2 of the fastest card out there in a system dont you think we are a bit CPU limited?

ATI's chips are not nearly as fast as 2 highend cards, all it is is CPU limitations and also unequal playing field (different HW).

-Kevin

So how do you know how efficient sli is? I agree one would assume greater results than this, but maybe there are still major bugs to work outl

I already said in my first post that i think there are also bugs that need to be worked out. But the fact is it is CPU limited. 1 can be CPU limited in a lot of cases... 2 is most definately.

-Kevin

You're not CPU limited if the XTPE is scoring higher. It's clearly a combination of already poor 6800 3dmark05 performance with some bug in SLI.

Look at the system specs of both? P4 3.4Ghz vs fx-53 oc'd to hell and back. The XTPE required dry ice as well, so its not like you can have it running 24/7. Its interesting, but not something that you can stick on your box or use in any real way. The nv system is something you could do out of the box (once the SLI connector is available), on an good, but not top of the line stock CPU.

I wouldnt mind the dry ice and massive OC, but I'd at least like to have the other components equal.

 

[Drayvn]

Originally posted by: DAPUNISHER

Originally posted by: Drayvn
This i think helps the argument that SLI might have a few bugs in it... maybe, dunno...

I think that is the likely answer, it's just not ready for primetime tech yet.

Maybe it might be down to the balancing of the load for each card, maybe ones doing to much work, effectively bringing down the score?