As you can see, the FPS drops in CellFactor were far more drastic when there was no PhysX PPU installed. This leads us to believe that the PhysX card plays a very large role in maintaining frame rates during intense CellFactor action. The average framerates for each test also support this conclusion, as the PhysX card afforded us with a solid 11% performance increase. Also, the first FPS drop on the graph is representative of the large initial explosion that can be seen in the attached video.

To view videos of the CellFactor gameplay/test procedure, download the following file: shoes_physx_cellfactor.zip (14.9MB, .wmv)

To further test the card with CellFactor, we ran a custom Performance Monitor log to keep track of CPU usage during our tests. Since the same visual effects are rendered by the graphics card on the screen whether or not the PhysX card is installed, the only possible place performance could be affected is on the Physics side. Therefore, the CPU must be doing substantially less work when the PhysX acceleration is enabled. Here are our results:



As you can see, the PhysX PPU offloaded quite a bit of calculations from the CPU. This was expected, as stated above. We found the average CPU usage decrease to be about 21.3%. This amount of CPU usage reduction would be expected to produce pretty significant performance increases, and it did, as shown above.

Conclusions
From these results it is clear that in certain games the PhysX PPU will be more effective than in others. Our results in GRAW did not produce any tangible performance increases, in fact we found it to decrease overall system performance while increasing visual grandeur. In CellFactor on the other hand, we found that the ASUS PhysX P1 physics card offloads as much as 21% of processor time from the CPU, which leads to an overall performance increase of just over 11%. Also, an apology to the developers of CellFactor is due, as in our PhysX article we devalued CellFactor as non-mainstream game, when in fact it is very fun to play once you learn the controls and has potential to be a widely played game.

However, this is probably the only game that has physics effects occurring almost constantly on such a large scale. As such, it is probably the only game that can exhibit performance increases of this magnitude. With this in mind, we can’t really change our verdict about PhysX technology – even with these numbers, it is still too early for us to advocate that gamers purchase a PhysX card. These results do, however, indicate that progression is occurring on a day to day basis. The time is almost right, it’s just not right now.

View the complete ASUS PhysX P1 review here.