While I really empathize with you guys, I've got to make a few points:
1) VIA's "official" ATA-100 support is still beta in the drivers, and definitely don't count the BIOS out here, either. Remember back to VIA and their initial ATA-66 support? It sucked big time for everyone but Asus, with drivers having little effect on performance. A few BIOS updates later, ATA-66 performance was normal for everyone. Same thing happened with my KT7 - HDTach's burst max for my IBM 75GXP drive initially was in the 45K range, a few BIOS updates later I check again and suddenly it's up to 59K, where it should be. Never changed the drivers, only the BIOS. Also, if you stop and think about it, MSI got their 686B boards out a lot faster than pretty much everyone else. Perhaps they skipped/skimmed the BIOS and driver validation process?
2) As for Tom - well, I wouldn't put a whole lot of stock in Tom's conclusions, either. His reviews may be OK at times, but he sometimes goes off the deep end when it comes to conclusions. Tom blamed VIA's initial ATA-66 problems on hardware back then too, when it was a BIOS/driver issue. Look at the chart on
this page. HDTach can benchmark a burst rate on 59K MB/sec on my 686A southbridge so I don't know where he got 30K MB/sec as the maximum for VIA's 686A. Actually, now I think about it, that number looks like the number he got from the first revision of the 686A with its initial BIOS and drivers. This was the also the same story with memory performance. Remember the standard story on how VIA memory performance on the Apollo Pro 133A sucked big time? Then the Asus P3V4X comes out and is within spitting distance of the BX. Seems VIA never told anyone about memory interleaving, IOQ, etc. and Asus was the first to figure it out. Again, a few BIOS versions later and most major system board manufacturers had BIOSes to boost memory performance.
3) Minor improvements aren't necessarily for the better. This whole thread exemplifies that.
4) As for VIA vs. Intel, VIA takes it a bit more of a hit because support for their chips isn't well integrated into the OS. There really wasn't an excuse for having to have a VIA patch for Win2K, and I believe that was Microsoft's fault. VIA's problem seems to be with communication and documentation. They apparently don't communicate with their OEMs what the capabilities of their products are, because it's usually Asus that comes out with an optimal product and then everyone else finds out what Asus tweaked. As a VIA board owner, there doesn't seem to be an official line on when to install 4-in-1s or in which order to install them, or which subset of the 4-in-1s is needed with which OS. VIA also does stuff like evolve from the KX133 to the KT133 but not change the chip identifier and again apparently doesn't tell anyone. Intel isn't perfect either because they've had their own driver problems as witnessed by anyone who's tried the wrong version of their ATA/100 drivers for the 8xx chipsets. Life was a lot easier driver-wise in the good ol' days with no AGP or UltraDMA.
5) Abit seems to take it on the chin a lot, and I've personally come to believe there's an Abit Curse
on people like Red Dawn and jonnyGURU. Their $25 fee for warranty RMAs is really ridiculous, but I've had no problems with my KT7 at all. I went from a Soyo BX to this KT7 non-RAID and it's been rock solid. I've got a Duron 700 running at 978MHz (8.5x115), a Monster MX300/Vortex 2 sound card, and a Geforce2 GTS running with AGP4x and sidebanding
and fast writes, and I've had no problems at all.
6) Even the best system board manufacturers don't bat 1000 all the time. What separates the top from the bottom is how they handle problems like this. I sure hope MSI does something for you guys.
Facebook
Twitter