PhysX performance in Borderlands 2

[Magic Carpet]

Is it true, that the older Geforce 400/500 series offer better PhysX performance in games than Kepler?

 

[BallaTheFeared]

Add in even more for GK110 "Kepler" and all signs point to "no".

 

[Carfax83]

Is it true, that the older Geforce 400/500 series offer better PhysX performance in games than Kepler?

No it's not true. But if you really want good PhysX performance, it's always best to get a dedicated PhysX card of some kind. A GTX 460 is probably the best bang for the buck PhysX card, or a GTX 650 ti if you want a Kepler based GPU.

 

[Magic Carpet]

You see.

Hello,

My name is Ray from NVIDIA Customer Care. We've been working with Gear Box Software on reports of performance issue with Borderland 2. We didn't get much specific on your particular issue except that performance is better on your 500 series than your GTX670? At 32 FPS that is within what we'd expect from a GTX670 since the GTX680 has a min FPS in the area of around 35 FPS. The latest 320.00 driver [http://www.geforce.com/drivers/beta-legacy] might improve FPS slightly assuming you're running PhysX High with a relatively low graphics setting. I discussed this issue with our development team and in certain conditions the 600 series may under-perform the equivalent 500 series cards since Kepler was highly optimized for graphics and does not perform as well on compute. At higher resolutions and graphics settings the GTX 670 will outperform the 570. Thousand cuts being a particularly heavy PhysX area and depending on which weapons are being fired (PhysX weapons effects) have a very heavy load. If your PC has 2 PCI-E X16 slots then the best configuration for PhysX games is to use the GTX670 for graphics and use the 5XX card as a dedicated PhysX GPU. That is assuming that the power supply can run both cards at the same time. The added compute capability improves the performance significantly and the task switching overhead of transitioning from graphics to compute is eliminated.


Best regards,
-Ray

Source.

Of course, I have no idea, how this is accurate as I neither own 670 or 570, but I am too having lag issues (in certain areas) with 660 in BL2 using latest drivers. Either the game's not well optimized, or am I asking too much from 660 @ 1050p? Thing is, neither CPU/GPU are fully stressed in those situations (I run 2nd monitor with monitor tools). It's just software not working properly :/

 

[Carfax83]

You see.

Source.

Of course, I have no idea, how this is accurate as I neither own 670 or 570, but I am too having lag issues (in certain areas) with 660 in BL2 using latest drivers. Either the game's not well optimized, or am I asking too much from 660 @ 1050p? Thing is, neither CPU/GPU are fully stressed in those situations (I run 2nd monitor with monitor tools). It's just software not working properly :/

Looks like Balla and I looked at the same source (PhysX Mark) and came to the same conclusion, which I guess isn't completely correct. Perhaps the Fermi derivatives do have some advantage..

At any rate, a 660 isn't a particularly strong GPU for intensive gaming, so adding PhysX to it's burden may not be a good idea. So either get a cheap dedicated PhysX card like a GTX 460, or get a faster graphics card; preferably 670 and above.

Also, post your specs, as the problem could lay elsewhere.

 

[BallaTheFeared]

You see.

Source.

Of course, I have no idea, how this is accurate as I neither own 670 or 570, but I am too having lag issues (in certain areas) with 660 in BL2 using latest drivers. Either the game's not well optimized, or am I asking too much from 660 @ 1050p? Thing is, neither CPU/GPU are fully stressed in those situations (I run 2nd monitor with monitor tools). It's just software not working properly :/

What CPU do you have?

http://1pcent.com/?p=135


GTX 680
2012-10-01 17:16:17 – Borderlands2
Frames: 127101 – Time: 1211191ms – Avg: 104.939 – Min: 0 – Max: 189
GTX 680 + GTX 580 SC (dedicated PhysX)
2012-10-01 01:18:30 – Borderlands2
Frames: 96895 – Time: 768648ms – Avg: 126.059 – Min: 13 – Max: 239
GTX 680 + GT 640 (dedicated PhysX)
2012-10-01 09:01:23 – Borderlands2
Frames: 420723 – Time: 3230437ms – Avg: 130.237 – Min: 23 – Max: 279


The validity seems suspect, but it's a data point.

Can you link your gpu usage and cpu usage charts just before, and during problematic areas - with vsync off.