When we published our review of the ASUS PhysX P1 physics card on Saturday, we only were able to test the card on one game, Tom Clancy’s Ghost Recon: Advanced Warfighter. We believe our reasons for choosing that game were certainly founded. It is, after all the only somewhat mainstream game that supports AGEIA’s PhysX PPU technology. However, as our tests and experience showed, GRAW is not a very good exhibitor of the PhysX technology. What we really wanted to test was a game called CellFactor, a futuristic Psy-FPS game that takes physics calculations to a level unseen on any current game, demo or otherwise.

Well, shortly after publishing the ASUS PhysX P1 review, we came across a method to play CellFactor without PhysX technology. The game enables it by default and does not let you toggle it off in the settings. We were able to work around this by “hacking” the game. The process is really not complicated at all, but we will not reveal just how we did it for liability reasons. We are sure that resourceful readers can figure it out on their own. So then as promised, here are our results from testing CellFactor with the PhysX card.

Test
As we had expected, CellFactor is a far better game to test the PhysX PPU on than GRAW, for two main reasons. First of all, the sheer scale of physics collisions that take place in a normal session of CellFactor is far beyond any sequence of events we could possibly dream up in GRAW. This means that the limits of the PhysX PPU will be tested to a greater extent than with GRAW. The second reason is that nearly every effect in the game is constant, whether or not you have the PhysX acceleration enabled. This means that, unlike GRAW, an explosion in CellFactor will be just as visually impressive, and graphically intense, whether or not you have the PhysX PPU installed. Therefore in this scenario, the PhysX card will indeed be handling the physics calculations that would otherwise be performed by the CPU.

Our testing method was very similar to what was seen in our GRAW test, we tried to replicate a sequence of events as accurately as possible in order to get a good benchmark for the PhysX performance. Two short videos of our procedure are included below.