A lot of hardware review sites place great emphasis on awarding marks for “stability “. How they can determine this from one sample I’ll never know, but they do. You the consumer should remember that even if a manufacturer achieves tolerance levels of the 95th percentile (or better), what this means in practice is that 1 person in 20 of you is gonna get burned.
Some sites also “test ” stability by looping various forms of software. Benchmark or gaming software is popular. If the machine crashes during this test, then they ascribe the fault (usually) to the motherboard. I say to this – possibly. I have seen this written about looping Winstone tests. I have also seen Winstone crash when testing for this article – but when the “crash ” happened, the Winstone user interface was looking for a disk in the A: (floppy) drive! How or why this happened I don’t know, but Ziff-Davis acknowledges a bug in the Winstone UI software, and has a patch for it. So was the crash due to a sub-standard motherboard, or “buggy ” software? Notice I didn’t tell you which board “crashed “, so now you know my feelings on the matter.
I also use Unreal (the game) as a great memory stress benchmark. If you overclock the FSB or tweak your memory timings beyond recognition, then I’ve found this a very useful test. If the memory can’t handle the settings you’ve chosen, Unreal’s timedemo will soon let you know. If a game crashes, is that the fault of the motherboard, or the memory? It would be good to know what crashed, when it crashed, what were the BIOS settings and what error was returned.
Having said that I can make some interesting comments about living with both motherboards.
The Abit has the VIA KT133A chipset, and there is one confirmed bug with it. If you have two hard drives on separate IDE channels (not as a master and slave set up on one channel), then copying large files from one drive to another will cause the PC to freeze. VIA have acknowledged this bug and released a patch. The patch comes with the VIA 4in1 driver package version 4.31 or later. Use it. Also, there are persistent rumours of difficulty with SoundBlaster Live! cards and crackling sound. I have not experienced this myself, but I didn’t specifically test for it. Another issue to consider. VIA’s official take on this is that it is a manufacturing tolerance issue, and you should take it up with the motherboard manufacturer.
And yet something else to make you paranoid, read what drivers you should install very carefully. Many people go out and get the latest 4in1’s and install them without thought. VIA IDE drivers (up until 4.29) should not be used with Windows 2000 – VIA said so. When I used the 4.25a’s and the 4.26B’s, using DMA mode for ATAPI drives (CD ROM, DVD etc) was almost guaranteed to produce corrupt files. No problems with the Microsoft default drivers which interestingly, default to PIO mode. ALi also make this recommendation on their website (use the Microsoft default W2K drivers). I have not had any issues under Windows 2000 with the VIA 4.29 or 4.31 drivers though.
IWill isn’t off the hook though. They also have a Windows 2000 issue and it’s a goody. When I benchmarked, I installed Windows 2000 on the 15GB 75GXP (with the latest BIOS) no problem. For my use, I use a 41GB 60GXP. What I do is partition the drive into four – go here to find out why. One partition each for Win98SE, Win2000, Linux (Mandrake) and data/applications. Guess what, Windows 2000 will not install on the KA266 using this partitioning scheme! The install crashes when it tries to partition the drive or check the partition after creating it. Sometimes it will partition OK and crash at the end of the format process. But crash it will. Windows 98SE and Linux (Mandrake 8.0) and Windows NT 4.0 all install without a problem using this scheme! Windows 2000 will install on a 15GB 75GXP using this partitioning scheme! This problem drove me nuts! The solution was to install NT4.0 and install Windows 2000 over the top. Sometimes updating the BIOS isn’t the cure all, and I tried a lot. I should point out that the BIOS version used is the latest official (not beta) BIOS on the IWill web site.
I’m inclined to take a full point off of the KT7A and half of one off of the KA266 for these “features “…
Don’t get me wrong, these bugs aren’t life shattering, and other motherboards and chip manufacturer’s have their fair share, but that’s beyond the scope of this article. When testing and set up correctly, I found both boards to be outstandingly stable (especially considering the aggressive performance options chosen).
|Score:||KT7A: 6.5||KA266: 6.5|
Other (random) thoughts
The type of application that wasn’t tested that may show DDR memory in a good light is multimedia – video, sound, photo manipulation. Perhaps in a follow up article when I can examine the software needed to properly evaluate the performance.
Overclocking. I didn’t get around to it, and it’s not as simple as upping the FSB until she blows and backing off a couple of steps… This might also make it into a follow-up.
Another area where DDR memory is supposed to shine is when used in conjunction with a chip that supports data pre-fetch, like the AthlonMP (and its desktop equivalent) and Pentium 4. Time will tell on this issue, but if it proves to be true, then taking DDR over SDRAM now might not be such a bad idea…
Be the first to discuss this article!