GTX 480 and GTX 470 Reviews

[trinibwoy]

Oh, as you've benched using FRAPS, maybe you could post some charts showing framerate fluctuation among cards tested during tests, mainly the GTX 470 SLI and HD 5850 CFX. That would be great!

Yes! HardwareCanucks puts out some of the best reviews around and that would take them up a notch. Since they're FRAPS fiends anyway it should be trivial to capture and graph the framerate over time.

[tajoh111]

lol wut

http://www.anandtech.com/show/2877/12

Its never slower, but it usually the same speed. But occasionally its a tiny bit faster.

[onethreehill]

Three Overclocked HD 5870s vs GTX 480
http://www.extremetech.com/article2/...2362685,00.asp

[ajaidev]

I want to see Quad CF/4 way 5870 "stock" vs Tri GTX 480 "stock"

Any review did that and very interesting results thanks OTH specially the gigabyte card.

[eric66]

Three Overclocked HD 5870s vs GTX 480
http://www.extremetech.com/article2/...2362685,00.asp

impressive graphs to say at least btw fermi really has huge problems with high res pretty much dies at 2560

[japamd]

For the most part, most reviews point to 5850CF being slightly faster than the 5970.

A few have shown it, but not the most AFAIK.

Anyway, look at these charts from hardwarecanucks review:

In BF BC2, 5850 CFX is 10.39% ahead of 5970.



In Dirt2, 5850 CFX is 19.20% ahead of 5970.



IMHO, that's very weird.

[overclock]

GTX 480 (825/1100 MHZ) vs. HD 5870 (975/1300 MHz), Overclocked Performance Analysis, Part 2
http://alienbabeltech.com/main/?p=16997

Would someone know this site ? The article is really interesting but I wonder if the site is reliable.

[iTravis]

A few have shown it, but not the most AFAIK.

Anyway, look at these charts from hardwarecanucks review:

In BF BC2, 5850 CFX is 10.39% ahead of 5970.


IMHO, that's very weird.

I don't know how they test it but my 5970 at stock scores 60fps average using FRAPS when I play at max settings 4xAA 16xAF @2560x1600.

[Chickenfeed]

I don't know how they test it but my 5970 at stock scores 60fps average using FRAPS when I play at max settings 4xAA 16xAF @2560x1600.

Yeah without context, comparisons are pointless given the huge number of variables in testing.

[iTravis]

Yeah without context, comparisons are pointless given the huge number of variables in testing.

QFT, I did notice the framerates are inconsistent for different maps in BC2, considering the game is still buggy and need to be fixed.

[japamd]

Would someone know this site ? The article is really interesting but I wonder if the site is reliable.

The results from that review looks flawed for the HD5870, they pushed memory oc too hard IMO and memory EDC seems to be holding back performance badly.

[SKYMTL]

For some reason, the HD 5970 really suffers in DX11 games. I know that the HD 5850 CF and HD 5870 CF had some good increases through the 1.0 Profile and post 10.1 drivers but I did not see any difference with the HD 5970.

[onethreehill]

ASUS GTX 470 Fermi Review


http://www.techpowerup.com/reviews/A...Force_GTX_470/

[damha]

Three Overclocked HD 5870s vs GTX 480
http://www.extremetech.com/article2/...2362685,00.asp

This is the first time I am seeing this



Where is the tessellation advantage?

[highoctane]

This is the first time I am seeing this

Where is the tessellation advantage?

Memory advantage:

(The latter, however, pulls ahead on the 2,560-by-1,600 version for the same reason as before; it has a bigger frame buffer than even the Nvidia card.)

[damha]

I saw that, still didn't explain it.

I think I figured it out: they are running v1.0 of the demo. That's silly.

[SKYMTL]

"Interesting" results to say the least. Even with the v1.0 of the benchmark, I don't think I ever saw results that low at 2560 x 1600 without AA enabled. :S

[damha]

"Interesting" results to say the least. Even with the v1.0 of the benchmark, I don't think I ever saw results that low at 2560 x 1600 without AA enabled. :S

They had two sets of runs, DX10 and DX11. Apparently thats the DX11 run. DX10 run was at about 20FPS. In your tests you had a much faster CPU clock.

[onethreehill]

ZOTAC GTX 470 - tested in SLI and compared to HD 5870 CFX
http://www.hexus.net/content/item.php?item=24263

GTX 480 Tested in 3 Way SLI and compared to HD 5870 Triple CFX
http://www.hardware.info/nl-NL/artic...GTX_480_test/1

[Olivon]

ZOTAC GTX 470 - tested in SLI and compared to HD 5870 CFX
http://www.hexus.net/content/item.php?item=24263

GTX 480 Tested in 3 Way SLI and compared to HD 5870 Triple CFX
http://www.hardware.info/nl-NL/artic...GTX_480_test/1

Thanks Adrian

[SKYMTL]

They had two sets of runs, DX10 and DX11. Apparently thats the DX11 run. DX10 run was at about 20FPS. In your tests you had a much faster CPU clock.

CPU clocks have very little impact at 2560 x 1600.

[damha]

CPU clocks have very little impact at 2560 x 1600.

You might be surprised to know that is not always true, so I am counting on this one being the oddball.

[SKYMTL]

You might be surprised to know that is not always true, so I am counting on this one being the oddball.

To the tune of ~100% difference in framerates? I have seen MAYBE a 10% change at high resolutions in even CPU bound games like Dragon Age and Left 4 Dead @ 2560.

I don't usually question results but those are on the extreme side...

[bamtan2]

so overclocked but noisy gtx 470 =< gtx 480 >= overclocked 5870

sounds (not literally) like a good place for the gtx 470 to be

too bad every review site is dropping the ball and not comparing these things directly? I guess you have to give them time to figure out how to overclock something.

[Pudi]

http://www.youtube.com/watch?v=vY8up...cc&feature=sub

Running EVGA's GTX 480 SC's in Quad SLi through a Heaven 2.0 Tessellation EXTREME benchmark.
1500W peak usage!