LGabrielPhoto
2[H]4U
- Joined
- Jan 5, 2006
- Messages
- 3,240
I guess it depends who you ask.
Follow along with the video below to see how to install our site as a web app on your home screen.
Note: This feature may not be available in some browsers.
Is a 8800GS on a X8 PCIE slot good enough?
Funny you mention this because I currently have a 9800 GTX+ and I am about to exchange it for a GTX260 (even trade...long story). I need to drive two hours to do so an I am trying decided if its worth it!
Let me see what I can try over the weekend to put and end to this discussion.
Key word there...Let me see what I can try over the weekend to put and end to this discussion.
Actually more like everybody, I would say YOU and a few others as there are another big bunch of guys including me, that think otherwise.Everybody knows that 9800+ cards are more than enough for physx
Lets stay on the topic, shall we?
Here you go, all tests were done with stock clocks on CPU, GPU and dedicated PhysX GPU.
This is an actual nVidia PhysX benchmark called "Star Tales Benchmark" from an Asian game "Nurien". Keep in mind this game was developed on a mature Unreal engine, closely with coordination with nVidia, the people that actually know what they are doing. Not slap-together trough emails like Cryostasis game by devs in Ukraine...
If pictures too to large for you to view hold Ctrl and scroll down with your scroll wheel
Graphics and Physics on 280 GTX @ stock speeds:
Graphics on 280 GTX @ stock speeds, Phisics on 9600 GT @ stock speeds:
Will there be any difference in PhysX processing between those 3 cards, or just get the cheapest?
Hmm, those results with the 295 show some serious gain over the GTX280 + 9600GT combo. Looks like that second massive GPU on the 295 is even better for PhysX than the 9600GT, as I figuredI matched your settings...
Q6600@ 3.81GHz
(1) 295 for Graphics set to C=602, SH=1458, M=1188
(1) 280 operating in Dedicated PhysX Mode set to C=756, SH=1512, M=1269.
*snip*
At 1920X1200 Max Settings... (With Motion Blur)
*snip*
from testing on my system, oc seems to matter very little for ppu. shaders are where its at.
I matched your settings...
Q6600@ 3.81GHz
(1) 295 for Graphics set to C=602, SH=1458, M=1188
(1) 280 operating in Dedicated PhysX Mode set to C=756, SH=1512, M=1269.
http://img530.imageshack.us/img530/3771/st1280.jpg
http://img530.imageshack.us/img530/5400/st1280bench.jpg
At 1920X1200 Max Settings... (With Motion Blur)
http://img141.imageshack.us/img141/8900/st1920.jpg
http://img140.imageshack.us/img140/5238/st1920bench.jpg
Hmm, those results with the 295 show some serious gain over the GTX280 + 9600GT combo. Looks like that second massive GPU on the 295 is even better for PhysX than the 9600GT, as I figured
uhmm vlad your not really nice in your replies , but anyway, the GTS250 comes with PhysX support right?
Well I should be getting a GTX260 in the next few weeks (dell) (or I find one at BB) and will post back the demo avgs. and then campare it against the 9600gso.
in the cryostasis tech demo and nvidia demos, the 260 will crush the gso. those demos benefit from high shader counts.
And the rest benefit from?
All I did was post a +1 like it's done by some many users so often...wow people is getting quite sensitive here.
So is the GTS250 on par with a 9800gtx+ or bit better? I've read in such places (mags and other forums) that if the 9800gtx+ was to be named into a GTX form, it'd be a GTS240.?? <?? Why has nvidia screwed up the naming scheme?
So how does this all compare with the 50 dollar dedicated Phys-X card I can buy from Ebay?
Think I found the test. It's the Firing Squad's Mirror's Edge review
http://www.firingsquad.com/hardware/mirrors_edge_physx_performance/page7.asp
According to them there ain't no difference between 9600 and 9800 for PhysX. Though, it's nice to see that one vid card + 2nd for PhysX gets as good result as SLI setup
Will have to get me Mirror's Edge and run my tests with/without my 9800