GTX580 official benchmarks

nest2001

n00b
Joined
Oct 15, 2004
Messages
31
gtx580D_copia.jpg

gtx580C_copia.jpg

gtx580B.jpg

gtx580A_copia.jpg
 
Seems the "less then 20% in games vs. 480" rumor was true. :)

Would be funny if reviewers were now benchmarking the cards up against OC variants of GTX 480, like Nvidia asked reviewers to do in the 6800 reviews. :p
 
Those charts are completely misleading. They make a 20% lead look like the card is going twice as fast. LOL.
I also liked how they compared gtx 580 sli vs radeon 5870 CF when they know that CF scaling sucks. Nice touch, Nvidia!
 
Those charts are completely misleading. They make a 20% lead look like the card is going twice as fast. LOL.
I also liked how they compared gtx 580 sli vs radeon 5870 CF when they know that CF scaling sucks. Nice touch, Nvidia!

LOL! And not to mention the total absence of AMD cards in the acoustics and performance per watt tests... :p

Don't you just love it when the PR monkeys from AMD and Nvidia puts out their official slides on how they view the world of graphics? :D
 
God I hate douche bags who make these charts start at something other than 0. Nvidia and AMD both does this shit, its just insulting. It serves no purpose than trying to mislead someone, which is lame in the first place, but secondly who the hell do they thing is that retarded exactly? Gah.
 
God I hate douche bags who make these charts start at something other than 0. Nvidia and AMD both does this shit, its just insulting. It serves no purpose than trying to mislead someone, which is lame in the first place, but secondly who the hell do they thing is that retarded exactly? Gah.


its annoying but its the marketing standard. been this way for ever. by now everyone should be able to read these things. they really aren't fooling anyone except the moron's that cant even read the charts in the first place.
 
It's for the investors.

And both AMD and NVIDIA do this. They have to bring to light the benefits of the new tech in the best-looking way possible.

These are not fake FWIW.

Plus, they have to pit this new GPU against what is OUT right now that is also a "Halo" single-GPU video card, so yeah, the 5870 has a target on its back.
 
It's for the investors.

And both AMD and NVIDIA do this. They have to bring to light the benefits of the new tech in the best-looking way possible.

These are not fake FWIW.

Plus, they have to pit this new GPU against what is OUT right now that is also a "Halo" single-GPU video card, so yeah, the 5870 has a target on its back.


yep. and also you really don't want to pit it up against a card that's going to compete with it in any way unless you know for sure the card your marketing can significantly beat that card in the real world.
 
I lol'd at the confidential stamped at the bottoms of every chart.
 
Those charts are completely misleading. They make a 20% lead look like the card is going twice as fast. LOL.
I also liked how they compared gtx 580 sli vs radeon 5870 CF when they know that CF scaling sucks. Nice touch, Nvidia!

If those charts are anything like the AMD ones for the 6850 and 6870, the 580 is only about 10-15% faster.......Here's a comparison:

http://img545.imageshack.us/img545/9959/174837ef5sngnbwu5bg1n5.jpg

Also, keep in mind that 6870 is estimated to be 16% faster than 460 1gb at stock.

http://techpowerup.com/reviews/Powercolor/HD_6850_PCS_Plus/28.html
 
So the Nvidia reviewers guide specifically asks Reviewes to use factory OC'd versions of the 480 to compare "real world performance" correct?

Somehow I doubt it.
 
If these are fake the person did a damn good job. Starting at 0.8 and positioning the graph and games as such. Also added in the tessellation graphs nice touch. Marketing team would be proud. But if you really look at it it shows nothing over 20% improvement for a single GTX 580 over the 480. If true this card should be called GTX 485 Emergency Edition FTW.
 
Guys the charts are not that hard to read. The card is being compared to the 5870 directly.. (Lame) Where the 5870 = 1, each line above 5870 is a 20% performance boost vs over the 5870 or other solutions.

Ignore the length of the bar, but pay attention to the data itself

So with the exception of Heaven (where each line is 10% faster) and the acoustic measurements (where each line up is 5 dba louder)

My take aways are;

gtx580D_copia.jpg

* at full (extreme) load the GTX 580 will be 47 dba loud which is 6 dba lower than 480 and not bad at all See ^^^

gtx580C_copia.jpg

* 40% better performance per watt over GTX 480 during the most power demanding scenarios See ^^^

gtx580B.jpg

* SLi 580 is anywhere from 18-120% faster than 5870 crossfire X in benchmarks (synthetic, canned/realworld ?) See ^^^

gtx580A_copia.jpg



* GTX 580 is anywhere from 18-120% faster than a 5870 in benchmarks (synthetic, canned/realworld ?)
 
Last edited:
Overall, I'd say the 580 is about 14-16% better.
Edit my math was wrong, Stupid confusing charts!!! LOL
Your not reading that chart correctly, I think your looking into the length of the bars and not the data

In most scenarios it's over 20% faster than GTX 480 (re-read the top of my other post before)

The card is either at or over 20% faster than GTX 480 in;

Batman Arkham Asylum
Starcraft 2
3dmark Vantage
Crysis Warhead
Hawx
Just Cause 2
Alien Vs Predator
Battleforge
Battlefield Bad Company 2
Dirt 2
Lost Planet 2
Metro 2033
Stalker COP
Stone Giant
Unigen Heaven

Only is less than 20% faster than GTX 480 in

COD MW2
Resident Evil 5
World in Conflict
 
Last edited:
Please Re-read the chart, look at the data not the length of the bars Each line up is a 20% boost over the other card

Well, here are my estimates based on what was presented in the last comparison chart. I've placed the numerical data to the right of each title.

The numerical advantage of the 580 over the 480 is as follows:

Batman Arkham Asylum: .2
Starcraft 2: .4 and 1/2
Crysis Warhead: .1 and 1/2 (?)
Hawx: .2 and 3/4's
Just Cause .2 and 1/2
Alien Vs Predator: .1 and 1/2
Battleforge: .2 and 1/2
Battlefield Bad Company 2: .1 and 1/2
Dirt 2: .2 and 1/4
Lost Planet 2: .4
Metro 2033: .2
Stalker COP: about .2 and 1/2
Stone Giant: .3 and 1/2
WIC: somewhere between 3/4's to .1
RE5: .1/4
MW2: .1 and 1/2

By comparison, the 6870 vs 460 1gb chart rates the 6870 as follows:

Batman: AA : . 2 and 1/2
Borderlands: .5 and 1/2
Crysis: .4
Crysis WH: .2 and 1/2
Dirt 2: .1 and 1/2
Quake Wars: .2 and 3/4's
L4D2: .3 and 1/4's
Metro 2033: about .3
Stalker COP: .3
RE5: .3
HAWX: .1 and 1/2
UT3: .1 and 3/4's
WIC: .1
Wolfenstein: .2
FC2: .1/2
BF: BC2: .3 and 1/2
Riddick: .1 and 3/4's


Now, comparing those two estimates, how much faster would you say the 6870 is over the 460 1gb? Also, keep in mind that Techpowerup rates the 6870 as being 15% better than the 460.

but man in other forums they are laughing and saying that most of you guys can't read the charts.

Links? I liked to view others opinion on this subject.
 
Well, here are my estimates based on what was presented in the last comparison chart. I've placed the numerical data to the right of each title.

The numerical advantage of the 580 over the 480 is as follows:

Batman Arkham Asylum: .2
Starcraft 2: .4 and 1/2
Crysis Warhead: .1 and 1/2 (?)
Hawx: .2 and 3/4's
Just Cause .2 and 1/2
Alien Vs Predator: .1 and 1/2
Battleforge: .2 and 1/2
Battlefield Bad Company 2: .1 and 1/2
Dirt 2: .2 and 1/4
Lost Planet 2: .4
Metro 2033: .2
Stalker COP: about .2 and 1/2
Stone Giant: .3 and 1/2
WIC: somewhere between 3/4's to .1
RE5: .1/4
MW2: .1 and 1/2

By comparison, the 6870 vs 460 1gb chart rates the 6870 as follows:

Batman: AA : . 2 and 1/2
Borderlands: .5 and 1/2
Crysis: .4
Crysis WH: .2 and 1/2
Dirt 2: .1 and 1/2
Quake Wars: .2 and 3/4's
L4D2: .3 and 1/4's
Metro 2033: about .3
Stalker COP: .3
RE5: .3
HAWX: .1 and 1/2
UT3: .1 and 3/4's
WIC: .1
Wolfenstein: .2
FC2: .1/2
BF: BC2: .3 and 1/2
Riddick: .1 and 3/4's


Now, comparing those two estimates, how much faster would you say the 6870 is over the 460 1gb? Also, keep in mind that Techpowerup rates the 6870 as being 15% better than the 460.



Links? I liked to view others opinion on this subject.

Hefly's never mind, my math was wrong and I created a cluster fuck basically. Everything in the chart is in correlation to 5870 so 480/580 can't be compared 1:1.
 
I'm guessing it'll perform like an overclocked GTX480. I won't really be impressed until I see how the 6970 compares.
 
Why is everyone always shocked and appalled when a company spins performance charts to make themselves look better? Have they not seen it before? Every company in almost every industry does this.
 
They should list fps too, if ATI is getting 10fps and Nvidia is getting 15fps as certain settings, you can claim a 50% improvement but it's still not playable. Reduce eye-candy to playable settings and it'll be like 50 and 55 fps compared, a small 10% improvement.
 
its only misleading to start at a number other than zero if you dont do that zig-zag line thing... i hope someone knows what im talking about. haha

which they didnt do. someone should edit it in to make it more accurate
 
Back
Top