VIA Issues – IDE and SoundBlaster Live
You may have read reports recently stating that there are issues with the VIA 686B Southbridge chip when using two ATA/100 hard disk with both set as Masters on the Primary and Secondary IDE ports. These reports also have indicated that this may involve the Creative SoundBlaster Live (SBL for short) or the SBL Value (SBLV).
The first reports seemed to come from the Register (http://www.theregister.co.uk/) – a publication viewed by some as a ‘tabloid’ due to it’s sensationalism and quite often a lack of hard facts to back up their ‘news’ – but in this case pointed to an article located here: http://home.tiscalinet.de/au-ja/review-kt133a-1-en.html. This article did a pretty good job of pointing out what they had found, but still left quite a few questions unanswered. There have also been numerous posts on a number of newsgroups and sites around the Internet discussing the topic. What I haven’t seen is any data to back up the original article, but have noticed (and participated in) some heated discussions on the topic. One trend I have seen is that most are just passing judgment without having all the necessary information. Quite frankly I don’t feel that without a background in mainboard and/or chipset design we just can’t place the blame, particularly without all the data needed to do so. But we can look and see if it is truly an issue and what configurations cause it, also what might cure it.
I must admit that I was skeptical when I first read the article. I sell a lot of VIA based mainboards using the 686B Southbridge, along with Western Digital ATA/100 Hard Disk and Creative SBLV cards, and had never heard of or seen any such thing. So I decided to check it out myself. I took a mainboard that uses the VIA KT133A and 686B chips and ran a series of tests to see if I could re-create the errors noted.
- I used a Soyo K7VTA Pro Mainboard with its latest BIOS (2AA2), Athlon 1.2GHz 266MHz FSB CPU, 128MB PC133, GeForce II GTS AGP card, D-Link NIC and one Western Digital 200BB ATA/100 7200 RPM hard disk as Master on the primary IDE port with a BTC ATA/33 CDROM as slave on the secondary IDE port. I then installed Win 98 SE, the latest VIA 4-in-1 driver pack (4.30) and the latest Nvidia drivers (dated 12/4/00). DMA was enabled for each drive. I then installed the two Winstone 2001 tests and ran each so they would repeat 5 times, installing the test programs each time – very extensive use of the hard disk. I also Zipped 3 large files (30MB to 90MB in size). There were No errors at all.
- Next I installed a SoundBlaster Live Value and the latest Liveware 3.0, and ran the test as above. No errors at all..
- Then I removed the SBLV, installed a second WD 200BB hard disk as Master on the secondary IDE port, set so DMA was enabled and moved the CDROM drive to Slave on the primary IDE port. I then ran the Winstone tests from the primary WD 200BB, but the actual tests were run on the secondary WD 200BB. This allows the data to be read from one drive, expanded and then the programs to be installed and run on the other drive. I also transferred the Zipped files from one drive to the other a total of 5 times. Again, there were no errors at all.
- Now, I re-installed the SBLV and tested as #3. Surprise! Now the system would either give me a hard lock-up or blue screen anytime there was any heavy disk activity between the two hard disks.
- To double-check my results, I then removed the SBLV and re-tested with each drive as a Master. Again, no errors.
- To see if it was an issue with ATA/100 mode I used WD’s utility to set each drive for ATA/66 mode and re-ran the test with the SBLV installed. The system again had disk errors.
- I then connected both drives to the primary IDE port, Master & Slave with the SBLV installed. Surprise again! I also had errors – not as often as with the drives on different ports, but I did have them.
- So as a final test I set PCI Delay Transaction to disabled and ran as #7. No errors this time.
So what did I find out? That if I used a Creative SoundBlaster Live Value on this Motherboard and two ATA/100 HDDs, the system was unstable and had disk errors. As long as the SBLV was removed there were no errors and the system was stable. Even under Win98…. Not just Win2K as some have suggested. Not even if connected to different ports.
What did I try to see if I could solve the problem? I tried a number of BIOS settings, the only one that helped was setting PCI Delay Transaction to disabled. But it did not solve the problem if the drives were connected to different ports, it just made the errors less frequent (though it did seem to run error-free if they were connected to the same port). It is interesting to note that setting is enabled for PCI ver. 2.1 and the SBLV is a PCI 2.1 card.
So what is the cause? Is it VIA’s fault? Or Creative’s? Hardware, BIOS or driver level? Could it have to do with PCI ver. 2.1? Or is just one of those incompatibilities that you just can’t plan on? I really have no idea, nor the technical background and data needed to make a judgement. But all does seem to be all related to the Creative SoundBlaster Live. No issues with the PCI NIC installed, though it’s on the PCI bus just like the SBLV (but is a PCI 2.0 card).
A couple of other notes: I did test with just one mainboard, just one chipset, just one Operating System, just one brand and model of hard disk. Results may vary with other components – so this was not an in depth test by any means, but enough to confirm there are issues.
It will be interesting to follow this over time and see not only what the ‘cure’ will be, but also if the true cause can be confirmed. I also plan on re-running the test with a 686A Southbridge and the new 8233 used with the Apollo Pro266 along with an Intel 815EP and the ALi MAGiK1 chipsets. Windows 2K and maybe ME will also need looking at. I wonder if I’ll see any problems with other Southbridges or chipsets, and the Creative SoundBlaster Live?
Be the first to discuss this article!