GTX470 benchmarks / comparison Tweet

[pixelrush]

Hey there crazy

I have just got a GTX460/2gb. it replaced a GTS250/1gb.
I 've been trying it out for size and I think 2gb is very useful.
Its actually probably more space than I usually need cos my renders are industrial design stuff and machinery with so not many textures but I thought the extra space was too good to pass up for just a few more $. I anticipate for most stuff I do I will need less than 1gb - the 1gb I had wasnt too limiting.
I also needed something which will fit in my pc case and run off my 550W PS.
I was aware that Octane rendering for hours is pretty hard on the card temp wise and also needed something fairly quiet. A 460 seemed to suit my needs best.
It gives a decent speed up to what I had before and makes changes to settings and navigating with higher resolutions 'ok' whereas before it was not really practical.
Here are some numbers for apportioning the total space I have been working out for myself:
4.5m poly[840mb] + A2@300ppi with narrow margins[655mb] + texture/hdri[290mb] + reserved space[260mb] = ~2gb
8m poly[1510mb] + A4@300ppi with narrow margins[155mb] + textures/hdri[100mb] + reserved space[260mb] = ~2gb
I have 8gb of ram and this is fully used loading 8m poly.I think if you had >2gb card you would want 12gb or more ram.
I think the most obvious benefit of 2gb is that it allows a decent number of polys and a large image size together. If you are a texture freak you will enjoy the extra too.
If you need a 3gb Tesla you are a pretty heavy duty user with arch viz needs I think.
6gb is probably for people who want to put in trees and grass in..not sure what the needs are there - not my area of work..
We havent heard anyone on the forums talking about top end requirements.
I would be interested to hear if anyone is finding 3gb too small.

The thing is that atm only 224 out of 336 cores of the 460 are doing anything for the performance you mention. Currently the performance of 2x460 are about the same as a single 480.
Cuda 3.2 due soon is supposed to fix that idle core issue + there is some possibility that some maths functions are working noticeably faster in cuda 3.2 for Fermi cards.
It remains to be seem if/how this will benefit Octane but it could be a decent speed up is coming, perhaps greater than 1.5x if we are lucky so 2x460 would then outpace a single 480.

[radiance]

switching cuda versions is always a disaster upon release, nvidia just changes too much.
so expect problems with a beta with cuda 3.2

also, we cannot develop release a cuda 3.2 octane render without it being released officially (it's in RC phase)
Just a bit more patience please, we are awaiting nvidia's release.

Radiance

[pixelrush]

I appreciate change breaks things but I can handle some discomfort if it means more working cores. I'm looking forward to the various improvements in v6

[buggy]

This is my first post on the forums. I just got Octane today and installed a Palit GTX 460 Sonic Overclocker Edition card with 2MB ram. I don't have room for a second card in the box so my monitor is connected to the new card.
Software: Octane v1.02 b2 - 64bit

MACHINE: AMD Phenom X4 9500 QuadCore
PC RAM: 6 GB
OS: Vista Premium 64bit
--------
GPU: Nvidia Palit 460 GTX - 2048 MB RAM
GPU Clock: 700 MHz
Memory Clock: 1800 Mhz
Shader Clock: 1400 Mhz
GPU TEMP : 67°C

RENDERTIME: 29m 22sec

The render is averaging 5.33 megasamples and peaking around 5.42.

Does this look like a healthy performance?
The Vtune tool will allow me to increase the clock settings for GPU, Memory and Shader. Is this safe to do? How do I monitor it or know at what point it is running too hot?
I don't want to shorten the life of the card.

I'm excited about using Octane to improve my results!

Thank you!

[pixelrush]

Hi, welcome to the forums. I have the same card and I like it for Octane.
I am a little unhappy though with the way the fan shroud ventilates mostly into the case.
I did have much the same arrangement with a GTS250 but the 460 runs quite a bit hotter and a fair bit of heat rises from the back to my other card I use for display.
This is actually my main concern.
At the moment I have removed some slot covers and pushed in a spare ram fan I had to pull the hot air out from between the cards - and I can tell you it is definitely warm air coming out the back.
I have two fans in the front of my case but the air really wasnt circulating enough to the bottom or between cards
I experimented a bit and this arrangement seemed to work best in my case although I might change it yet..if anyone has a better idea I would appreciate them sharing it.
Pulling the heat off the back of the card was much more effective than trying to boost the air to its fan.
Although the 460 will typically overclock 25% I went for 12% core and 6% memory as a balance between a bit extra speed and looking after the components. I use MSI Afterburner which will draw you a graph of temp as well. Once you have it set up and you are happy it runs cool enough under load you shouldnt need to worry about it.
The Palit doesnt have memory chip cooling so I was playing more safe there.
At the moment with the o/c and 80% fan it runs about 68deg C and my other 58 in continuous use.
I would feel a little happier if I could get that down to say 65 and 55 but I dont think its too bad.
Because the other card isnt doing much I feel it ought to be ~50 but the rising heat and not so good air circulation are giving it a tougher time than it should. It only has a small fan so it is easily 'upset' by extra work.
If you have 67 deg I would say thats ok. If you o/c with fan control and its much the same I would say its fine.
I am pleased with the performance of the 460. I thought the GTS250 was ok for my needs but the 3x speed up is well worth having esp for larger resolutions. I 'll be chuffed if those other 112 cuda cores will come to life with cuda3.2. Man if you have 3x480 using Octane must be so sweet
Anyway hope that blah helps you.
GPU definitely makes unbiased rendering fun. Thats an attribute you cant really express in statistics.

[SamPage]

For those considering a GTX 400 series, be aware that NVidia appears to have crippled the transfer speeds on the card (specifically glReadPixels()), so if you get one or more of these cards to render on Octane, you are going to want a Quadro, or probably best an AMD/ATI Radeon or FireGL to actually drive the display that you are modeling with.

Sam

[GR1F1TH]

I've been testing out my 3x gtx465's these are the numbers I pulled from the benchmark scene.

DL=Direct Lighting PT=Pathtracing MS=Megasamples
1024x512
1GPU @ DL=5.9MS PT=1.64MS
DL load
86%
2GPU @ DL=10.2MS PT=3.08MS
DL load
82%
83%
3GPU @ DL=11.6MS PT=4.3MS
DL load PT load
71% 90%
63% 92%
63% 92%

1920x1080
1GPU @ DL=7.3MS
DL load
97%
2GPU @ DL=13.1MS
DL load
90%
93%
3GPU @ DL=18MS PT=4.78
DL load PTload
91% 89%
82% 94%
85% 96%

4096x3112
1GPU @ DL=9.2
load
99%

2GPU @ DL=17.7
DL load
96%
98%
3GPU @ DL=25.8MS PT=5.7MS
DL load PTload
96% 99%
91% 99%
94% 99%

I am pretty disappointed that the cards are not being utilized fully at lower resolutions in Direct Lighting. They're so low that the 3rd card is useless. In Path Tracing I'm pretty happy with the results in the 90's. Is something going on here in DL? I have not searched much in the forums but I would think that the cards would be fully used no matter the resolution. Does this have something to do with the cuda 3.2 update soon?

[bazuka]

people could any1 confirm to me whats faster (i know its a stupid question)

PALIT GeForce GTX 260 Sonic 216 SP 896MB
or
PALIT GeForce GTX 460 Sonic 1GB GDDR5

im planing to get Gtx 460 and i just want to make sure that im making a right choice

thx

[matej]

The 460 is a newer architecture and has 336 cuda cores (vs 216). Get the 460 without any doubts.