• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

Minimum for PhysX?

Chosen Juan

Limp Gawd
Joined
Mar 30, 2011
Messages
138
I was just wondering, what is the minimum card required for PhysX? I have an 8800GS lying around and just want to know if it's worth it to try it out for PhysX, mainly Batman: Arkham Asylum since it's on sale on Steam right now.

If it's not worth it then, oh well.

Thanks
 
Depends on the title, something like an 8800GC would be fine for most games, and for AA it should be fine. As you own all the hardware, it wouldn't hurt to try :D
 
Here is a post I made in the PhysX section which I believe you'll find helpful: http://hardforum.com/showpost.php?p=1037405239&postcount=4

Depends on the title, something like an 8800GC would be fine for most games, and for AA it should be fine. As you own all the hardware, it wouldn't hurt to try :D

Yeah, might as well give it a shot. Just be wary: the 8800GS will add ~100 watts to your rig's overall heat output, and if you place the card beneath the R6950 it will make the card heat up more than if there was nothing blocking it.

Case in point: my top 6950 runs 10* C hotter than the bottom one, even though the top one's fan is running 20% faster and the card is using 1063 mV instead of 1093 mV on the bottom card. You can negate this issue by buying a ribbon cable and having one card or the other outside of the case, but that's another issue entirely. For now, you might as well try it and see if you like the trade-off.

Good luck.
 
Here is a post I made in the PhysX section which I believe you'll find helpful: http://hardforum.com/showpost.php?p=1037405239&postcount=4



Yeah, might as well give it a shot. Just be wary: the 8800GS will add ~100 watts to your rig's overall heat output, and if you place the card beneath the R6950 it will make the card heat up more than if there was nothing blocking it.

Case in point: my top 6950 runs 10* C hotter than the bottom one, even though the top one's fan is running 20% faster and the card is using 1063 mV instead of 1093 mV on the bottom card. You can negate this issue by buying a ribbon cable and having one card or the other outside of the case, but that's another issue entirely. For now, you might as well try it and see if you like the trade-off.

Good luck.

Would the 8800GS run fine at x4 mode? I could place it on the 3rd (most bottom) PCI Express slot which always runs at x4 so the 6950 stays at x16.
 
Would the 8800GS run fine at x4 mode? I could place it on the 3rd (most bottom) PCI Express slot which always runs at x4 so the 6950 stays at x16.

I'm pretty sure you'd get 97%+ of the maximum possible performance by using an x4 slot for PhysX, you can give it a try to find out.

Also, my example in the above post is with double-slot spacing, or cards in slots 1 and 4. Increasing the spacing more should reduce the top card's airflow penalty.
 
It's been proven with benchmarks that the faster your main GPU is, the faster your dedicated PhysX card has to be or else it will slow down the overall performance. What kind of bullshit is that? It's not like PhysX effects are getting better. In many cases you're better off just running PhysX on your main GPU. Or, you could just make a stand against the proprietary, unoptimized bullshit Nvidia puts out, and not use a second card.
 
I was just wondering, what is the minimum card required for PhysX? I have an 8800GS lying around and just want to know if it's worth it to try it out for PhysX, mainly Batman: Arkham Asylum since it's on sale on Steam right now.

If it's not worth it then, oh well.

Thanks

physx or no physx, batman aa is worth the price of entry on steam, so go buy it :)

and a 8800gs (comparable to a 9600gt) should work just fine with the game.

Would the 8800GS run fine at x4 mode? I could place it on the 3rd (most bottom) PCI Express slot which always runs at x4 so the 6950 stays at x16.

x4 should have no effect on performance with the 8800gs for gpu physx

It's been proven with benchmarks that the faster your main GPU is, the faster your dedicated PhysX card has to be or else it will slow down the overall performance. What kind of bullshit is that? It's not like PhysX effects are getting better. In many cases you're better off just running PhysX on your main GPU. Or, you could just make a stand against the proprietary, unoptimized bullshit Nvidia puts out, and not use a second card.

he has a 6950 so can't use gpu physx on that; and he has a 8800gs lying around so he may as well use it if he is gonna play batman.

your statement is mostly true, generally speaking - but in batman, not so much at all:

http://www.tomshardware.com/reviews/batman-arkham-asylum,2465-8.html

there is practically no difference (with a gtx 260 as primary) between using a 9600gt and a gtx 260 as a dedicated ppu in batman.

so yeah, don't think him using a card he already owns as a ppu will do much in sticking it to the man.
 
Last edited:
Thanks for the responses guys. Well, in addition to AA I picked up the THQ Pack as well which has Metro 2033. It has hardware PhysX if I'm not mistaken. Would that be a game where an 8800GS would slow down my overall performance?
 
Thanks for the responses guys. Well, in addition to AA I picked up the THQ Pack as well which has Metro 2033. It has hardware PhysX if I'm not mistaken. Would that be a game where an 8800GS would slow down my overall performance?
how can the 8800gs slow you down? that would only be possible if you were comparing it to running it on another Nvidia gpu. your only other option is your cpu so the 8800gs, in general, should handle it better than that.

did you even read anything I said to you in the other thread though? its not really worth the hassle of sticking in another video card and using hacked drivers just for physx. this is especially so in Metro 2033 where even if you watch a comparison video the visual difference is minimal.

and again try running it on your cpu first because Sandy Bridge quads can handle hardware physx pretty darn good. just look at the Metro 2033 with physx on cpu results. http://www.xbitlabs.com/articles/cpu/display/core-i7-2600k-990x_9.html#sect2
 
Yeah PhysX is almost useless, And the sandy bridge can do it for Metro im not sure if it can for AA.
 
how can the 8800gs slow you down? that would only be possible if you were comparing it to running it on another Nvidia gpu. your only other option is your cpu so the 8800gs, in general, should handle it better than that.

did you even read anything I said to you in the other thread though? its not really worth the hassle of sticking in another video card and using hacked drivers just for physx. this is especially so in Metro 2033 where even if you watch a comparison video the visual difference is minimal.

and again try running it on your cpu first because Sandy Bridge quads can handle hardware physx pretty darn good. just look at the Metro 2033 with physx on cpu results. http://www.xbitlabs.com/articles/cpu/display/core-i7-2600k-990x_9.html#sect2

in the xbitlabs review, isn't that just advanced physx enabled? the reason i ask is because at pcgameshardware, their chart shows a difference in performance between advanced physx and gpu-physx:

http://www.pcgameshardware.com/aid,...marks-with-DirectX-11-and-GPU-Physx/Practice/

the tested 5850 showed no difference in performance between advanced physx on and off. that probably explains the minimal visual difference.

also, found this relevant thread:

http://hardforum.com/showthread.php?t=1522512&highlight=metro+physx

i think the op should try it out and report back. but it was definitely worth it for batman aa, imo; and maybe batman ac too if that uses the same engine.

i know last year when i played mafia 2, a dedicated physx card helped in that game a whole lot with gpu physx enabled. just my two cents.
 
in the xbitlabs review, isn't that just advanced physx enabled? the reason i ask is because at pcgameshardware, their chart shows a difference in performance between advanced physx and gpu-physx:

http://www.pcgameshardware.com/aid,...marks-with-DirectX-11-and-GPU-Physx/Practice/

the tested 5850 showed no difference in performance between advanced physx on and off. that probably explains the minimal visual difference. found this relevant thread:

http://hardforum.com/showthread.php?t=1522512&highlight=metro+physx

i think the op should try it out and report back. but it was definitely worth it for batman aa, imo; and maybe batman ac too if that uses the same engine.

i know last year when i played mafia 2, a dedicated physx card helped in that game a whole lot with gpu physx enabled. just my two cents.
when you turn on advanced physx in Metro 2033 it uses additional hardware level effects. if you have an Nvidia card you can have that run on the cpu if you have it set to do so in the Nvidia control panel. otherwise it will just run on the gpu of course. with an AMD card if you select advanced physx it will run those effects on the cpu only.

I have no idea what is going on with their results there. I tested it before in the benchmark and there was a bigger framerate loss running the effects on my cpu than on my gpu.
 
when you turn on advanced physx in Metro 2033 it uses additional hardware level effects. if you have an Nvidia card you can have that run on the cpu if you have it set to do so in the Nvidia control panel. otherwise it will just run on the gpu of course. with an AMD card if you select advanced physx it will run those effects on the cpu only.

I have no idea what is going on with their results there. I tested it before in the benchmark and there was a bigger framerate loss running the effects on my cpu than on my gpu.

okay, just wondering.

just noticed the xbitlabs has unusual results for metro - isn't that game a gpu heavy game? why is there such a large performance differential between the tested cpus? since according to legionhardware, there should be little to no difference:

http://www.legionhardware.com/articles_pages/metro_2033_performance_guide,9.html

i don't get it...
 
how can the 8800gs slow you down? that would only be possible if you were comparing it to running it on another Nvidia gpu. your only other option is your cpu so the 8800gs, in general, should handle it better than that.

did you even read anything I said to you in the other thread though? its not really worth the hassle of sticking in another video card and using hacked drivers just for physx. this is especially so in Metro 2033 where even if you watch a comparison video the visual difference is minimal.

and again try running it on your cpu first because Sandy Bridge quads can handle hardware physx pretty darn good. just look at the Metro 2033 with physx on cpu results. http://www.xbitlabs.com/articles/cpu/display/core-i7-2600k-990x_9.html#sect2

I enabled PhysX on both Normal and High on Batman: AA and my framerate drops like no other. I guess I'll break out the 8800GS tomorrow to see how it is, but I honestly don't know if it's worth having it for just one or two games.
 
when playing metro with my old 5870 there were fewer minimum frame dips when i threw in a 9800gt and hacked physx. did make for a better gaming experience.
 
I was gonna use a GTX260 as a physx card to complement a GTX470, but I don't have a power supply that's big enough to handle the load and after thinking about it I don't think the added power draw would be worth it in the end. That could all change if a good game that supports physx comes out that I really like.
 
Well, I might try it out but I'm not sure how to go about doing it. Do I still need hacked/old nVidia drivers to get it working?

So pretty much: install the 8800GS, install the drivers for it, and I should be good to go right? I saw a guide on another website but it seemed like it was too much work, requiring another monitor etc.
 
Back
Top