![[H]ardOCP](/images/blank.gif){kind=small}
- Date:
- Monday , October 19, 2009
- Author:
- Mark Warner
- Editor:
- Brent Justice
- Google +1
Batman: Arkham Asylum & PhysX Gameplay Review
In Batman: Arkham Asylum, the Dark Knight pummels an endless stream of thugs, but will the game deliver the same two-fisted brutality to your video card? Find out as we examine gameplay performance and image quality on nine of today's video cards. Extensive testing of NVIDIA's PhysX and what it means to your gameplay experience!
CPU and Memory Usage
To measure CPU utilization and System Memory consumption, we used the Microsoft Performance monitor, which is a standard component of Microsoft Windows, located at "C:\Windows\System32\Perfmon.exe" for Windows 7. The CPU and Memory usage tests are performed just like the video card performance tests, except that FRAPS is not necessary. The Performance Monitor software automatically measures and graphs the data. We used the same gameplay run-through for CPU and Memory usage as we used for the video card tests.
For the sake of equality, all of these tests were performed at 1920x1200 with 8X MSAA, 16X AF, and maximum in-game graphics options, with the exception of the PhysX setting. These first two tests were performed without PhysX enabled at all, and then we explored the effect of PhysX on CPU and Memory usage in the last three tests. We used 1920x1200 instead of 2560x1600 because of an apparent bug with NVIDIA hardware on our test platform that causes catastrophic performance degradation at 2560x1600 with PhysX at the maximum setting.
For the NVIDIA testing, we used the GeForce GTX 285. For the AMD testing, we use the ATI Radeon HD 5870. These video cards represent the very best in single-GPU performance from each manufacturer.
CPU and Memory Usage Without PhysX
These tests were performed at 1920x1200 with 8X MSAA, 16X AF, and all in-game options enabled, except for PhysX, which was disabled for this test. For good measure, we made sure to disable PhysX acceleration on the GPU in the NVIDIA control panel as well as the in-game PhysX option.
ATI Radeon HD 5870
Using AMD’s ATI Radeon HD 5870, we found that core 3 had the highest consistent usage, averaging 55% and peaking at about 94%. Overall, the CPU load was spread fairly well across all four cores, as there was not a single core that was idling or one that was completely loaded all of the time. On average, we saw 38% CPU utilization on our quad-core processor, with a peak of about 60%. This seems to indicate that there should not be a performance hit on dual-core processors as opposed to quad-core.
On the memory side of things, the test system showed 3141MB of available system memory sitting at the desktop. After we loaded up the game, that decreased to about 2509MB average, with a low of 2437MB available. That means that Windows (and its services and drivers) was using 955MB of memory by itself. At its peak, the game used about 704MB of memory in this test.
NVIDIA GeForce GTX 285
With the GeForce GTX 285, we saw small changes. Core 0 was the most loaded, at 65% average and 100% peak. The CPU as a whole saw about the same usage, coming in at 37% average utilization with 63% as a peak.
With the ForceWare 191.07 installed, the system used 914MB of memory sitting at the desktop, which is 41MB less than with the Radeon HD 5870. At peak usage, Batman: Arkham Asylum used 713MB of system memory, which is 9MB more than with the ATI video card.
CPU and Memory Usage With PhysX
These tests were performed with the NVIDIA GeForce GTX 285 at 1920x1200 with 8X MSAA, 16X AF, and all in-game options enabled , except for PhysX, which we toggled according to whichever test is indicated on the graph.
NVIDIA GeForce GTX 285 with PhysX Enabled on the CPU
This test was performed with the GTX 285 installed and PhysX acceleration enabled on the CPU. The game was configured for the fullest PhysX option available.
With the GeForce GTX 285 installed and PhysX running on the test system’s quad-core CPU, core 1 saw the most usage, averaging 44%, with a peak of about 97%. The CPU as a whole averaged 39% usage, peaking up to about 71%.
At the desktop, the system was consuming 973MB with GPU PhysX acceleration disabled. At its peak, the game used 766MB, which is 53MB more than with PhysX disabled completely.
NVIDIA GeForce GTX 285 with PhysX Enabled on the GPU
This test was performed with the GTX 285 installed and PhysX acceleration enabled on the GPU. The game was configured for the fullest PhysX option available.
With graphics and PhysX running on the same GTX 285 GPU, our CPU usage changed hardly at all. Core 0 was the most heavily loaded, at an average of about 51% and a peak utilization of 97%. Across all four cores, we saw an average utilization of 35%, and a peak of 71%. That is only 4% lower on average than with PhysX running entirely on the CPU.
With GPU PhysX enabled, the system used 874MB of memory sitting at the desktop, which is 99MB less than with PhysX running on the CPU. At peak load, we saw the game’s memory usage rocket up to 1382MB, an 80% increase in memory utilization.
NVIDIA GeForce GTX 285 with PhysX Enabled on a dedicated GTS 250
This test was performed with the GTX 285 and a GTS 250 installed and PhysX acceleration enabled and configured to run solely on the GTS 250. The game was configured for the fullest PhysX option available.
With a GeForce GTX 285 handling graphics and a GTS 250 handling PhysX acceleration, core 0 was again the most heavily loaded, averaging 62% with a peak of 100%. Across the whole CPU, average usage was 44% with a peak of 78%.
At the desktop, the system consumed 954MB of memory with both of these video cards installed and configured. The game’s peak load condition consumed approximately 1099MB of memory. That is 283MB less than with the GTX 285 handling both graphics and PhysX, and 333MB more than with the GTX 285 handling graphics and the CPU handling physics.