|MSI Radeon HD 4850 Video Card R4850-512M|
|Reviews - Featured Reviews: Video Cards|
|Written by Olin Coles|
|Tuesday, 14 October 2008|
Page 7 of 13
Crysis Benchmark Results
Crysis uses a new graphics engine: the CryENGINE2, which is the successor to Far Cry's CryENGINE. CryENGINE2 is among the first engines to use the Direct3D 10 (DirectX10) framework of Windows Vista, but can also run using DirectX9, both on Vista and Windows XP.
Roy Taylor, Vice President of Content Relations at NVIDIA, has spoken on the subject of the engine's complexity, stating that Crysis has over a million lines of code, 1GB of texture data, and 85,000 shaders. To get the most out of modern multicore processor architectures, CPU intensive subsystems of CryENGINE 2 such as physics, networking and sound, have been re-written to support multi-threading.
Crysis offers an in-game benchmark tool, which is similar to World in Conflict. This short test does place some high amounts of stress on a graphics card, since there are so many landscape features rendered. For benchmarking purposes, Crysis can mean trouble as it places a high demand on both GPU and CPU resources. Benchmark Reviews uses the Crysis Benchmark Tool by Mad Boris to test frame rates in batches, which allows the results of many tests to be averaged.
The very first thing we discovered in the low-resolution tests was how seemingly poor both of our multi-GPU products performed. The Gigabyte GeForce 9800 GX2 was matched in average frame rate by the MASI Radeon HD 4850, and the GeForce 9800 GTX edged out the CrossFireX set of 4850's. To be fair, none of these video cards will probably ever realistically see this low resolution, so the performance only illustrates how high-end GPU power can be cut short if the monitor (resolution) doesn't match it.
Low-resolution testing allows the graphics processor to plateau maximum output performance, which thereby shifts demand onto the system components. At the lower resolutions Crysis will reflect the GPU's top-end speed in the composite score, indicating full-throttle performance with little load. This makes for a less GPU-dependant test environment, and is helpful in creating a baseline for measuring maximum output performance in the next few test results.
At the 1280x1024 resolution used by 19" monitors, our results show that performance is beginning to really drop despite the small difference is pixels drawn. In terms of general performance, all of these products maintain the same performance ratio as before, except for the 9800 GX2 which seems to hold its ground.
The CrossFireX set of HD 4850's is going to soon reach it's limit, as it is in last gear and the 9800 GX2 is still shifting up the tree. The NVIDIA GeForce GTX 280 Reference Product is outperformed by the overclocked ZOTAC GeForce GTX 280 AMP! Edition with a 11% margin; which goes to show how far a 100 MHz overclock will take the GTX200 GPU.
For widescreen users, our benchmarks below indicate that the ATI Radeon HD 4850 matches the performance of NVIDIA's GeForce 9800 GTX video card, although the 4850 stops delivering post-processing effects at 8x AA and the 9800 GTX can reach 32x AA if the application supports it.
A CrossFireX set of HD 4850's beats out the GeForce 9800 GX2 by almost 13% at this lower widescreen resolution, but they can't touch the GTX 280 series... yet. Testing in high-pressure Crysis also seems to have effected both of the GeForce GTX 280 products we've tested, which are barely ahead of the GeForce 9800 GX2 dual-GPU graphics card.
Heading into the 1920x1200 resolutions produced on the SOYO DYLM26E6 used for testing, Crysis forces 2.3 million pixels to be processed by our graphical test products. Despite what 3DMark06 has reported, the CrossFireX set of Radeon HD 4850's is not king; the GeForce GTX 280 series is. If only by a small difference, the overclocked ZOTAC GTX 280 enjoys a 5% (2 FPS) lead over the GeForce 9800 GX2, which also seems to keep its own 8% (3 FPS) lead over the CrossFireX set.
At our highest widescreen resolution, the overclocked Foxconn 9800GTX-512N performs the same as MSI's R4850. At the end of our Crysis testing, neither the GeForce 9800 GTX or CrossFireX HD 4850 set could touch the single ZOTAC GTX 280 AMP! Edition video card. But before we leave Crysis, I decided to include a look at post-processing performance with 4x AA enabled at the 1680x1050 resolution; which is really about the only resolution that AA is playable with these products.
Since NVIDIA has recently reduced the price of GeForce 9800 GTX products to compete with the HD 4850, there will be some intense fighting between these two products. My professional opinion is that if these two products shared the exact same price and I only wanted to buy just one of them, my money would go to the Radeon HD 4850 over the GeForce 9800 GTX. If you want excellent bang for the buck from a multi-GPU array, my money would go to CrossFireX because of performance and widespread compatibility.
Even with a decent dose of anti-aliasing added to Crysis (at 1680x1050), the performance is still relatively decent for all products. Our Island timedemo mixes some beach and water views so it's going to be on the high side of frame rates when compared to actual game play, but as you can see every product we've tested hovers near the 30 FPS barrier for playable frame rates. It's worth noting that the reference GTX 280 produces a 3% lead over the 9800 GX2, which is minimal at best, and the 14% lead that a ZOTAC-overclocked GTX 280 AMP! Edition can deliver may not make the best argument for price.
In our next section, Benchmark Reviews switches to video-output only benchmarking with Lightsmark 2007. Read on to see how a blended high-demand GPU test with low video frame buffer demand will impact our test products.