OTOY Forums

First legitimate review: http://www.hexus.net/content/item.php?item=21651&page=1

Conjecturing somewhat, GeForce GTX 480 is probably 75 per cent of the high-end GPU that was imagined by NVIDIA early last year. Our numbers show that NVIDIA's finest single-GPU card is, on average, around 20 per cent faster than AMD's Radeon HD 5870 1,024MB at a 2,560x1,600 resolution. GeForce GTX 480 is due to cost some 40 per cent more, so whilst the trade-off between extra expense and performance isn't ideal, it's not shockingly bad, either.

Not as bad as some had hoped afterall!

edit: hexus.net is down at the moment but someone managed to capture all the benchmarks:
http://www.computerbase.de/forum/showpost.php?p=7597425&postcount=3418

hmm, true or not.

http://www.slashgear.com/nvidia-geforce ... d-2278420/

its a fail project in my eyes

Did you tried it with raytracing?

even if fermi is slower than the latest ATI card in gaming / directX fixed function pipeline tests,
it will be totally different, and MUCH faster for cuda based raytracers due to the additional cache they have.

Radiance

[gk] wrote:the fact that this gfxcard needs over 300w is just plain silly, I have no desire to get that kind of powerwaster

Actually its closer to 250w, which is still kinda high - i'm going to wait for the November refresh which should solve the power consumption problem plus i'm waiting for a 3GB card

a review(its more games related):

http://www.fudzilla.com/content/view/18242/1/

We need a Octane render review on GTX 480, speed of rendering is important!

Hope fermi is much faster than it looks to seem?

Geforce GTX 480 / 1344,96 Gflops
Geforce GTX 470 / 1088,64 Gflops

Geforce GTX 285 / 1062,7 Gflops
Geforce GTX 275 / 1010,9 Gflops

(Radeon HD 5870 / 2720 Gflops)

GFLOPS is not everything

There's more space for kernels, and you can run multiple kernels at the same time
GFLOPS can be used to compare GPU only if they share the same architecture

Wich is not the case with the GF100 chips

I think I'll wait a bit before I buy one - right now I have a GF295 which is entirely sufficient for what I need it for. As far as increased Octane performance is concerned, I don't really care at the moment since it will be at least a couple of months before Octane gets all the features I need for production use. Plus the additional time will perhaps let nVidia work out some inevitable kinks and improve drivers. And who knows, maybe the price will drop a bit by then.

[gk] wrote:the fact that this gfxcard needs over 300w is just plain silly, I have no desire to get that kind of powerwaster

But think about it in Octane terms... imagine how much power you would need to run multiple cards to get the same CUDA core count. More processing power needs more power period. And think about how much you would save compared to having a render farm with multiple boxes to work with other render engines. The people I know who have render farms run up huge power bills.