Test System
| CPU: |
Intel Pentium IV 2.4C - 200MHz FSB
|
| Motherboard: |
DFI Infinity 865PE
|
MSI 865PE Neo2
|
Albatron PX865PE Lite |
| Memory: |
1GB Corsair XMS TwinX PC4000 Kit (2*512MB) |
| Memory Speed |
3-7-4-4 |
2.5-8-4-4 |
2.5-5-3-3 |
| Hard Drives : |
40GB Seagate ST340016A, 40GB Maxtor 34098H4, 2*80GB Maxtor 6Y080M0 SATA Hard Drives
|
| Video Card: |
Matrox Parhelia 128MB (200MHz/250MHz) |
| Operating System: |
Windows XP Pro SP1 Direct X 9 |
| Drivers: |
Parhelia 1.04.00.78 |
| Cooler: |
Vantec Thermalflow |
| Case: |
CoolerGuys Windtunnel IV |
| Power Supply: |
RaidMax 400Watt Power Supply |
| Direct X Benchmarks: |
Unreal Tournament 2003 (HardOCP software 2.1 - CPU Test) |
| OpenGL Benchmarks: |
Jedi Knight II Jedi Outcast (time demo) |
| Other Benchmarks |
VirtualDub 1.4.10 |
DivX 5.03 |
| |
Truespace 4.2 |
Pi Fast 4.2 |
| |
TMPGEnc Plus 2.59.47.155 |
Office Bench Online - HT Analysis |
The benchmarks were run as has been seen in previous reviews, but lets go over it again. For the Jedi Knight the test was run at 1024*768 with all the settings on high except for AA and ansiotropic filtering. Unreal Tournament 2003 was run using the HardOCP CPU test at 640*480 to cut the Parhelia from causing a bottleneck in this case.
VirtualDub was run using a 8555 frame DVD file converted to HuffYUV at 720*480 and then encoded the DivX 5.03, without the audio being encoded. Truespace was run using the settings mentioned previously, and rendered to 1600*1200 to this. TMPGEnc was run using the same video as the VirtualDub test, though in this case the audio was also encoded.
Pi Fast was run using 10000000 digits of Pi, Chudnovsky method, 1024 K FFT, and no disk memory, just as we at VL have done previously. Our new benchmark Office Bench was run at a resolution of 1600*1200@32bpp, with the printer used being a HP Color Laserjet 5/5PS print to file, printer. Now lets look at the gaming benchmarks.
Gaming Benchmarks
Jedi Knight has been around for quite awhile, and being based on the Quake III engine, it's going to be memory bandwidth limited. So how does the performance differ between the motherboards in this highly CPU limited test?
What can we see with this test? It seems that the DFI motherboard performs slightly better than the Albatron motherboard by just under 3%. It also performs better than the MSI 865PE that we have tested earlier, both in single channel and in dual channel mode. Do other games show the same performance difference as with Jedi Knight?
Unreal Tournament was released to much hype, and stressed both video cards and at 640*480 stressed the processor as well. HardOCP released a benchmark for the game, which has both video and CPU tests, and we will use version 2.1 of the benchmark in this test. So how well does this motherboard work?
Here in this game we see that the difference between the two motherboards follows the performance of the memory speeds. The MSI motherboard takes the lead in dual channel memory by 1% over the DFI motherboard. From the top of the results to the lowest results we see that there is only a 3.2% difference between them with the Albatron 848P leading in single channel mode. We've seen that in games the DFI motherboard performance is among the top of our small group though the results are very close so as to make the difference insignificant in the results. How does this motherboard perform with all our video encoding tests?
Video Encoding
Our video encoding test suite starts with the program XMpeg which we took the vob file that we mentioned and converted it, without audio to DivX 5.03 video using the SSE2 option in encoding the video. Since this is highly memory intensive and CPU intensive, how does DFI perform in this test?
Here in this test we see that this program favors dual channel memory. Dual channel memory gives you over 6% of an improvement over its single channel alternative. The difference between the two 865PE motherboards is extremely minute and non-existent. What happens when we take the same video in a lossless compression and encode it in VirtualDub into DivX 5.03?
Next Page
