AMD 760 vs. ALi MAGiK1 at 266MHz FSB and PC2100

Pages: 1 2

The Test

OK, now it’s time to see which is really faster and by how much. Since the 1.133GHz Athlon is the only speed 266MHz FSB CPU I’ve received so far, that’s the CPU I’ve used on both mainboards. As with the other tests already run I used 128MB of Micron CAS 2.5 PC2100 DDR SDRAM (set at CAS 2 for the test), an ATI Radeon 32MB DDR AGP video card, WD 200BB ATA/100 7200 RPM HD, had a PCI NIC installed, used Windows 98 SE with the latest drivers from AMD and ALi running at 1024×768/64K @ 85Hz. All tests were run 3 times and rejected if the scores were not within 3% of each other.

What I’ve done is list the two chipsets / mainboards side by side along with how much faster in percentage the AMD is compared to the ALi, and an average of the percentage column at the bottom.

Now I will admit that when running the tests and compiling the results I was a bit disappointed, but once I looked at the actual percentage results I felt much better. When I started to see the difference in most cases was only in the 1 to 2 percent range I knew that the difference was so slight you’d never even notice it.

Conclusion

A couple of interesting things caught my eye. Winstone and SYSmark, the two business application based benchmarks, don’t give the same percentage difference. Could it be due to one running just a single program at a time and pre-loading it into memory (SYSmark) and the other running with multiple windows open (Winstone)? Or is it due to the different programs used in the test suites? CPUmark 99 is very close (under 1%), but the CPU test for 3D Winbench and 3D Mark shows a larger spread. So where is the overall performance difference coming from (doesn’t appear to be CPU or memory related)? Looking at the graphics, 3D and video test results I almost wonder if it’s an issue with the AGP performance of the ALi that is not quite as good as the AMD, rather then being a memory timing issue.

So is an AMD 760 chipset faster than an ALi MAGiK1? Yes, at least with the mainboards I’ve tested. Could that change? Maybe a bit (in favor of the ALi) but I doubt it. Is one that much faster then the other? I would say no, since the average of my tests was 2.28% in favor of the AMD, and most were within 1~3% with the biggest spread was a bit under 7%.

Do I have reservations about recommending the ALi over the AMD? No, the performance difference is minimal. I expect to see a much better selection of ALi mainboards compared to the AMD, with better pricing and more features (especially for those that wish to overclock). I think sometimes you have to look at the ‘big picture’, not just the actual performance numbers. In reality, does it matter that one has a Winstone score of a half of a point higher when that is only about a 1% difference?

It will be interesting to revisit the two chipsets in a month or so and see if the results are the same as the products mature. Hopefully we will be able to include a VIA KT266 based Mainboard at the same time, and maybe even a chipset from SiS.

Pages: « Prev  1 2