We used several video benchmarking programs to test how the HD 4850 X2 performs. We used 3dMark 2006 and 3dMark Vantage to test the card’s synthetic performance. FRAPS was used to benchmark “real-world” frames-per-second in a variety of popular PC games. Our test setup consists of:
- Lian Li PC-V1200 Black Aluminum Case
- Intel Core2Duo E6600
- Zalman CNPS 9700 NT CPU Cooler
- Zalman ZM-MFC2 Multi Fan Controller
- Asus P5WDH Deluxe motherboard
- OCZ 8GB DDR2-800 Platinum Edition
- Asus Xonar D2X PCIe Sound Card
- Seagate Barracuda 7200.11 1TB Hard Drive
- Logitech Wave Keyboard
- SteelSeries Ikari Laser Mouse
- ThermalTake Toughpower 850W PSU
- Sapphire HD 4870 Toxic video card
- Sapphire HD 4850 X2 video card
- eVGA GeForce GTX 280 Video Card
Power Consumption
The power consumption of all of our video cards were measured using Zalman’s ZM-MFC2 Multi Fan Controller (reviewed here). We tested the cards at full idle and then under load (by letting 3dMark Vantage run). It took several seconds before each card hit its peak, which we then recorded as a final measurement. We discounted any measurement that didn’t last for more than a second, but most of the time the wattage went up and down like a bell curve as the graphics load increased and decreased accordingly.
| idle Watts | load Watts | |
| Sapphire 4870 Toxic | 227 | 309 |
| Sapphire 4850 X2 | 232 | 352 |
| Sapphire 4850 X1 | 238 | 295 |
| GeForce GTX 280 | 220 | 346 |
It is interesting to note that the GeForce GTX card is more power efficient when idle, but not by much. At full load, the GTX uses up as much power as the 4850 X2, and far more than the 4870 Toxic.
3dmark 06 Benchmarks
| 3dmark score | SM 2.0 Score | SM 3.0 Score | |
| 4850 X2 | 7698 | 10565 | 4243 |
| 4850 X1 | 5603 | 6300 | 4081 |
| 4870 Toxic | 6351 | 8658 | 3529 |
| GeForce GTX 280 | 6937 | 9722 | 4153 |
3dmark Vantage
3dMark Vantage goes beyond your normal video benchmark. It focuses on the two areas that really matter in a gaming PC seperately, namely the CPU and GPU. In previous benchmarks you got an overall score that could leave some questions up in the air as to how powerful your system really is. For example, our current test machine uses a relatively old Intel E6600 processor, and thus the overall score could make the video card appear not as powerful as it truly is. Now there are seperate scores for both the graphics processor and main processor.
3dMark Vantage has two GPU tests, each having a different emphasis on rendering. CPU tests are covered by performing a physics simulation and AI simulation
| 3dmark score | GPU score | CPU score | |
| 4850 X2 | 7698 | 10565 | 4243 |
| 4850 X1 | 5603 | 6300 | 4081 |
| 4870 Toxic | 6351 | 8658 | 3529 |
| GeForce GTX 280 | 6937 | 9722 | 4153 |
Far Cry 2 Benchmarks
Far Cry 2 uses a different engine than Far Cry and Crysis, namely because the engine was developed by Ubisoft rather than Crytek. There are not as many dynamic features of Crytek’s previous game (like real-time motion blur). The game is still gorgeous, and the engine supports a new dynamic gameplay element featuring fire.
Regardless of the engint, it appears that the developers haven’t gotten the memo about Crossfire and SLI. Far Cry 2 can be demanding, but currently the highest FPS number come from non-SLI setups. All of the cards in our lineup can play Far Cry 2 very well, but just don’t deliver the number we expect from a multi-GPU card.
| low | avg | high | |
| X2 1900×1200 | 26 | 32 | 43 |
| 4870 1900×1200 | 29 | 33 | 44 |
| X1 1900×1200 | 20 | 28 | 42 |
| X2 1280×720 | 30 | 61 | 82 |
| 4870 1280×720 | 33 | 40 | 53 |
| X1 1280×720 | 31 | 48 | 61 |

2 comments
r4i software
Best-performing card in the $300 price range four DVI outputs let you connect up to four LCDs.
The thing is, I know very little of power supplies. So my question to you is: what would be the best power supply for this graphics card? I read that it has to be at least 650W, but besided that, I don't know what to look for.
dsi r4
Comments are closed.