Nvidia GeForce GTX 500 Series | 580 570 560 | Reviews | General Discussion

[PhantomAU]

Originally Posted by snappy1

Tox was probably honest

Sorry, couldn't resist, we will let Tox explain if he wants to

http://www.shanebaxtor.com/category/nvidia/

btw shane... I get a lot of these errors on the page:

WordPress database error: [Table './shanebax_blogdb/wp_statpress' is marked as crashed and should be repaired]
DESCRIBE wp_statpress;

WordPress database error: [Table './shanebax_blogdb/wp_statpress' is marked as crashed and should be repaired]
SHOW INDEX FROM wp_statpress;

[flu!d]

um no, on average it is 15% faster. some games will be 50% some will be 20%. thats why you take averages.

On average, under peak conditions the 480 can be up to around 30% faster than the 5870.

?? I am aware of the meaning of the term average.

[Luke212]

Originally Posted by flu!d

?? I am aware of the meaning of the term average.

On average, under peak conditions the 5870 can be up to around 30% faster than the 5870.

ok.... whatever you are trying to say, the 480 is 15% faster than the 5870.

you could cherry pick a few games and average them and the 480 would be 70% faster, but that's hardly in the spirit of objectivity.

[flu!d]

ok.... whatever you are trying to say, the 480 is 15% faster than the 5870.

On average it's 15% faster....At peak it can be up to around 30% faster - Quite obviously.

Performance is generally measured at peaks, you don't say my car has a top speed of 75kp/h because that's what it pulls on average?!

I haven't seen a scenario yet where the 480 is 70% faster than the 5870.

[Autti]

Originally Posted by flu!d

Performance is generally measured at peaks

What? Since when are graphics cards evaluated at peak performance, that means cherrypicking one game.

[Luke212]

Originally Posted by flu!d

On average it's 15% faster....At peak it can be up to around 30% faster - Quite obviously.

Performance is generally measured at peaks, you don't say my car has a top speed of 75kp/h because that's what it pulls on average?!

I haven't seen a scenario yet where the 480 is 70% faster than the 5870.

you're probably more experienced at arguing here and im not trying to start one. i think we are actually saying the same thing so ill leave it at that.

[Mau1wurf1977]

Originally Posted by Alba

I get the same results in crysis, with the 5870 being a touch faster.

I just noticed that techpowerup says "Crysis" and Anandtech "Crysis Warhead". So maybe that's the reason?

Because in the Anandtech review at 1600 4xAA the 480 is almost twice as fast

http://www.anandtech.com/show/3987/a...range-market/8

Originally Posted by flu!d

I haven't seen a scenario yet where the 480 is 70% faster than the 5870.

Anything with heavy tessellation and the 480 can be twice as fast.

Now I got to hand it to Nvidia regarding tessellation...

I mean ATI has been promoting tessellation for many many years. And Nvidias first DX11 card blows ATIs tessellation out of the water

I see tessellation only becoming more important, which is good for anyone with a 400 or 500 series card.

Should be interesting to see how Cayman fares...

[flu!d]

Anything with heavy tessellation and the 480 can be twice as fast.

Interesting, I've never noticed this before.

Cheers Luke,

I wasn't trying to start an argument, just pointing out the fact that the 480 generally is quite a bit faster than the 5870, if the 580 is 15-20% faster 'on top' of the lead the 480 currently enjoys than the 6970 has quite a challenge ahead of it.

It's going to be interesting to see if the 6970 has the legs....

[mAJORD]

Originally Posted by flu!d

if the 580 is 15-20% faster 'on top' of the lead the 480 currently enjoys than the 6970 has quite a challenge ahead of it.

It's going to be interesting to see if the 6970 has the legs....

I don't think 30-35% over its predecessor will be too much of a challange for Cayman if the rumoured specs hold true.


By the way, it's either Peak or Average, not both. ..

It (480) is up to 30% faster maybe in the odd game.. 10% slower in another odd game, ~15% faster overall across a selection of games.. i.e and Average. No two ways about it really.

[Alba]

Not much in it percentage wise in vantage

Both 5870's and 480's OCed by 100MHz on the core.
480's OCed by 76MHz on the mem, 5870's at stock
Both runs done in the same PC.

3 x GTX480's @ 800/1000


3 x HD5870 @ 950/1200

[James086]

0.953v core voltage?

[TOX]

Originally Posted by Alba

Not much in it percentage wise in vantage

Both 5870's and 480's OCed by 100MHz on the core.
480's OCed by 76MHz on the mem, 5870's at stock
Both runs done in the same PC.

3 x GTX480's @ 800/1000
image


3 x HD5870 @ 950/1200
image

Is it possible you're hitting a CPU limitation?

I'm getting over P30k with HD 6870 CF OC, HD 5870 CFX OC and GTX 480 Tri-SLI should be getting more then +5k which is only about 15%.

You tried an X run.

That would remove any chance of a CPU limitation, it would also be interesting to see what happend with the scores as AMD are stronger at the higher resolution.

[Alba]

Originally Posted by TOX

Is it possible you're hitting a CPU limitation?

I'm getting over P30k with HD 6870 CF OC, HD 5870 CFX OC and GTX 480 Tri-SLI should be getting more then +5k which is only about 15%.

You tried an X run.

That would remove any chance of a CPU limitation, it would also be interesting to see what happend with the scores as AMD are stronger at the higher resolution.

Can't do a X run atm with the 480's as the screens are 1920x1080.
The CPU doesn't effect the GPU score much and the same CPU at same clocks is used in both, so it's a fair comparison.

Originally Posted by James086

0.953v core voltage?

That's just at idle, under load it goes up to 1.18v.

[DiGiTaL MoNkEY]

GTX 580

[Bertross]

interesting the min is worst then a 480.