GTX 580 benchmark

[BlueBread]

Hi,metalliandy.

Lets just hope that we don't encounter this limitation on the GTX 560 which is coming soon .
I say this because the design of the GTX 560 should be modified for greater efficiency and power than the currently released GTX 5xx.
In addition to this Nvidia could remove this bothering limitation that we all dislike by accident or intention . However, all his is based on my hopes, no facts, so lets hold our thumbs .

[SamPage]

Well, we just had a chance to very quickly bench a GTX 570 here on Rhino, and it has the same issues the 4xx had. On a standard Rhino benchmark (Holomark) the GTX 285 turned in 100 frames at 16.02 seconds, while the 570 took 26.05s.

Boooooo.

Sam

[metalliandy]

@BlueBread,
Sorry i have not responded earlier...i didnt get a notification email (that or i missed it :/)

@SamPage,
Well that sucks

Did anyone try the new beta drivers that came out today?

[gah5118]

I just ran the benchmark scene on my machine
gtx 580 3gb
i7 950
24gb ram

the results were
rendertime = 11m35s
20.27m/s on average
24-25 fps

as for running maya and rhino it handles fine so far. I was running a ati 6870 till about a few hours ago. I picked up the 580 for octane and a few other cuda based softwares.

hope this helps someone.

[asterix7]

Running Octane Benchmark on beta demo 2.46b on a Mac MBPro 17' with NVidia 8600M GT-512Mb
5300/16000 s/px after 2h07min, 0.86 fps - 0.52 Ms/sec
This card has 32 CUDA streams, core clock 475 Mhz shader clock 950 MHz
Theoretically this card should do 16x longer on CUDA , or maybe 32x longer considering it runs practically at 1/2 speed, then it still should take 7 hours.

How does shader and core clock of the 8600GT compare to Graphics (777 Mhz) and Processor clock (1544Mhz) of the GTX580

Would 2xGTX580 be redering in 1/2 the time ?

[Marcusin3D]

Could someone put here 580 bench. ?

Hi. My first post here

Finally done a big upgrade myself which inlcudes (after months of agonising over a quadro) a Palit GTX 580 3Gb.

First thing I tried was of course the benchmark and got:

16000/16000 s/px in 13 mins 21 seconds

Averaged 22.12 fps
Averaged 13.73 Ms/sec

Hope that helps

Marcus

[grimm]

Just to flesh out a comparison, here are the numbers for the benchmark at 16000 SPP on my Pallit 2GB GTX460 336 cores:

pathtracing - 1.73 MS/s 3.30 FPS Time - 1:29:54
PMC - 1.67 MS/s 3.18 FPS Time - 1:32:56

This is using 2.47 of course.

[karanis]

Using demo version of Octane,
On benchmark scene, with default settings, nothing touched, I got :

Zotac GTX580 (on current Slackware 64-bit with current dev_drivers and stable CUDA toolkit) : 00:10:42
Zotac 8800GT Amp (on 32-bit WinXP) : it`s been 01:24:30

Zotac GTX580 is installed on a Gigabyte UD7 with i7 and 2000Mhz 6 gig ram. *And I did my best on ventilation

hope this helps

but it is still slow.. wrt hardware costs

[9fly]

Hi Karanis,

Is your Zotac a AMP2 with 3GB version. How is the speed with other render so far. I'm looking for one pair of this but still consider with MSI Lightning Xtreme Edition cause this card only 2 slots profile. But Zotac is a bit better deal at this time.

[t_3]

Maybe would be helpfull if somebody could make some perf. summary results with at least 450, 460, 470, 480, 570, 580 and some quadro or tesla cards.

except quadro/tesla cards i've done exactly that:
http://www.refractivesoftware.com/forum ... f=9&t=7383
...