Summary: Today NVIDIA is releasing a new PhysX-enabled tech demo for Cryostasis, the upcoming survivor-horror game from 1C. With rich graphics sporting tens of thousands of simultaneous water particles, how well will the various GPUs hold up running PhysX in this test? Is PhysX even worth it? Find out inside this article!
The state of PhysX gaming today
Cryostasis
Described as a first-person survival horror shooter, Cryostasis is set in 1981 aboard the Russian nuclear icebreaker North Wind. You’ll play the role of Alexander Nesterov, a meteorologist that’s been dispatched to investigate what happened onboard the ship. Apparently the ship ran into trouble near the North Pole and it’s up to you to find out what happened to the ship’s crew – did they die due to the harsh arctic environment or was it something else?
Survival plays a huge role in the game. Not only must you survive against the ship’s crew, equally important is surviving the elements. With the game taking place in the North Pole, the environment is frigid. Since everything is frozen, you’ll have to watch the outside temperature as well as your own body temp. After all, if you get too cold, you die. Because of this, you’ll spend the game constantly looking for sources of heat (instead of featuring a traditional health meter, Cryostasis relies on a heat meter). Lights, pipes, fires, can all be potential sources of heat. As you heat up areas to stay warm, the ice turns into liquid and this is where PhysX kicks in. [image]
PhysX is used for fluid effects simulation in Cryostasis. More specifically, smoothed particle hydrodynamics (SPH). According to NVIDIA “Unlike a regular particle system which typically only accounts for collision, SPH computes viscosity and repulsion forces which make the particles splash and puddle as if they were part of a real fluid. Additional PhysX features include tarps, curtains, rag dolls, rigid bodies.” [image]
Today NVIDIA has released a tech demo showcasing the PhysX effects in Cryostasis. The techdemo runs you through a scripted sequence showcasing many of the game’s PhysX features. We’ll provide you with a quick pictorial walkthrough…
With the game taking place on a frozen ship with ice literally melting right in front of you, water plays a large role in Cryostasis. The techdemo really pushes the hardware, with the first sequence containing nearly 30,000 fluid particles that interact with the environment and other objects. The techdemo opens with a daunting character standing underneath a leaking pipe. Water literally rains down on him, with the water splashing realistically off his body: [image]
The water realistically pools around the character and collects around objects on the floor: [image]
In the next sequence, your character walks into a frozen room with a cloth hanging from the ceiling. The cloth is stiff from the ice on it. Your character then pushes a button, heating up the room, melting the ice, and allowing the cloth to flap around in the breeze from the fan: [image]
Next your character shoots an enemy. His lifeless body falls into a pool of water where water flows realistically around him: [image]
After turning on another heater, a hose shoots water down a flight of stairs. Looking closer you see that the water from the hose splashes up against the wall and pools on the ledge before falling down the staircase: [image]
At the bottom of the stairs the water particles sheet down the wall and splash off the steps before pooling at the foot of the stairs: [image]
Now its time to show off the game’s buoyancy effects. Barrels float in the water differently based on their weight, the deep pool of water ripples and splashes as objects are shot into the water: [image]
Next you walk into a room with objects on a table. Water interacts realistically with the objects, sheeting off the table and falling on the floor, where it pools into a puddle after falling off the table: [image]
In the final sequence, you walk into a room where a character is throwing barrels onto a large cloth tarp. The barrels bounce off the tarp before falling on the floor: [image]
Hardware requirements
In order to run the PhysX techdemo on the GPU, a GeForce 8-series or better card is required. However, if you don’t have a CUDA-compliant GeForce 8 or better graphics card, the techdemo will run PhysX on your system’s CPU. Therefore those of you with Radeon 3800/4800 series hardware can check out the techdemo if you’d like, although the game runs significantly slower with PhysX running off the CPU, we’ll provide exact benchmarks for those of you seeking specifics a little later in this article. Needless to say it isn’t a pleasant experience, and it’s exacerbated by numerous graphical artifacts and rendering errors we saw running Radeon 4800 hardware with Catalyst 8.12.
Intel Core i7 920 Overclocked to 3.3GHz Intel X58 Smackover Motherboard 3GB Qimonda DDR3-1066MHz NVIDIA GeForce 9500 GT DDR3 NVIDIA GeForce 9600 GSO 384MB NVIDIA GeForce 9600 GT NVIDIA GeForce 8800 GTS 640MB NVIDIA GeForce 8800 GTX NVIDIA GeForce 8800/9800 GT NVIDIA GeForce 9800 GTX+ NVIDIA GeForce GTX 280 NVIDIA GeForce GTX 260 w/216 shaders NVIDIA GeForce GTX 260 ForceWare 180.84 ATI Radeon HD 4850 512MB ATI Radeon HD 4870 1GB ATI Radeon HD 4870 X2 2GB Catalyst 8.12 300GB Western Digital Caviar SE Windows Vista Ultimate 64-bit w/Service Pack 1 Benchmarks
Cryostasis Techdemo
![]() Notes
As you can see, all of the high-end graphics cards run at 11 frames per second when PhysX is handled by the CPU. Our OC’ed Nehalem processor just can’t run the Cryostasis PhysX techdemo any faster regardless of the high-end GPU used. Nehalem actually runs the techdemo faster than the GeForce 9500 GT handling a mixture of graphics and PhysX.
Unfortunately due to time constraints we weren’t able to run SLI PhysX (where two cards are running together in SLI graphics mode with one card handling a mixture of graphics and PhysX) verus multi-GPU PhysX benchmarks (where one card is dedicated solely to graphics, and a second card dedicated solely to handling PhysX), but we did manage to get some good multi-GPU PhysX benchmarks, as well as determine a possible baseline for the minimum hardware you’ll want for dedicated GPU-based PhysX processing. We hooked up a 9800 GTX+ handling graphics with 9800 GTX+, 8800 GT and 9600 GT cards that were dedicated just for PhysX. Here are the results: ![]() As you can see, all three cards yield the same performance numbers with the 9800 GTX+, suggesting you don’t need anything better than a 9600 GT for handling dedicated PhysX processing. Let’s see what happens when we pair this 9600 GT with cards ranging from the 8800 GT up to the GTX 280: ![]() With the GeForce 9600 GT dedicated solely for PhysX processing, the primary GPU is able to focus exclusively on rendering graphics. We saw a nice performance increase as a result, with the GeForce GTX 200 and 9800 GTX+ cards seeing a performance improvement of 17-18%, while the 8800 GT’s performance improves by 25%. Not bad at all! We should note that we took another stab at running the 9600 GT in the secondary PCI Express graphics for PhysX with ATI’s Radeon HD 4870 handling primary graphics duties. Unfortunately however the NVIDIA graphics driver wouldn’t let us do this, so mix match PhysX appears to be limited to GeForce cards only (i.e. mixing a 9800 GT with a 9600 GT). UPDATE 2PM: We just heard back from NVIDIA regarding mixing and matching ATI and NVIDIA hardware. Unfortunately this isn't supported by Vista. The Vista driver model only supports one graphics driver at a time.
We’re not going to come to any final conclusions based on this techdemo though. As we stated earlier, there’s still a lot we simply don’t know about this game. For instance, since the techdemo doesn’t provide a setting for turning PhysX on or off, we don’t know how the two modes compare visually to one another. We have a feeling the PhysX version looks better, but how much better? Obviously we’d assume the PhysX version will have more fluid particles, but how far will 1C be willing to push things graphically? It’s okay to push 30,000 particles in a techdemo, but it’s highly doubtful they’ll go that far in the final game. The majority of gamers don’t have GeForce GTX 200 cards after all. Another point of concern is frame rates. Since the techdemo is based on a prerecorded sequence, we can’t get a feel for what an acceptable frame rate is. It’s possible 30 fps could be a playable frame rate, or 40-50 fps may be more ideal. It doesn’t look like Cryostasis will be a twitch shooter, so it’s highly doubtful more than 60 fps will be needed for an ideal gameplay experience. Our frame rates in the techdemo maxed out at 38 fps for the GeForce GTX 280 handling both graphics and physics duties, and 46 fps when the card was paired with a 9600 GT dedicated solely for physics. While neither of these frame rates are particularly high considering the resolution used (1600x1200), again keep in mind that techdemos tend to exaggerate things graphically so this probably isn’t representative of the final game. Besides, it’s prerelease code anyway. So overall the Cryostasis techdemo represents a neat little PhysX performance demo to play with, but nothing earth shattering just yet. The GeForce cards running PhysX delivered significantly faster performance than the Radeon hardware relying on the CPU – in our tests the 216-shader GTX 260 ran Cryostasis three times faster than Radeon 4870 – but those same GeForce boards ran just as fast as the Radeon cards when both were running CPU-based PhysX. Hopefully we’ll be able to report back to you with more concrete details on Cryostasis and its PhysX implementation once we get closer to the game’s final release next year. | ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| © Copyright 2003 FS Media, Inc. |