Potential Data Corruption Problem on Tyan Thunder MBs?

wezelboy asks: "I need independent verification of a data corruption problem with the Tyan Thunder LE mobo during DMA writes to an IDE drive. The problem manifests itself under heavy disk load in large files, does not change file size, and only corrupts 4 byte chunks. Thus, it is not easily spotted. We are running Red Hat 7.1, but this corruption may be reproducible on other operating systems." There's more detail of this problem in the article. Any of you Tyan Thunder motherboard owners care to help track this problem down?

"A simple way to verify this problem follows:

Using a Red Hat 7.1 install (2.4.2 kernel, but other 2.4 kernels exhibit this problem) and a file /tmp/x that is over 500MB in size, "cp /tmp/x /tmp/y", and compare the md5 checksums using md5sum. Keep doing this, deleting /tmp/y each time before you do.

/tmp/x and /tmp/y will continuously show the same size, but will sometimes have different md5 checksums. A diff between the files usually shows that 4 bytes are moved to another part of the file. You may have to go through a number of iterations before you see it. We have a more sophisticated test script available, just e-mail me at here and I'll send it to you.

Tyan claims that this problem is a limitation of the Serverworks LE chipset, however I have not been able to reproduce this problem on any of the Asus CUR-DLS motherboards (which use the same chipset!).

If anyone can reproduce this problem on a different Serverworks LE mobo, I'd like to hear about it. Thanks!"

Comments

I would say yes

[Outland Traveller]

I have a Tyan Thunder K7 Rev. B motherboard.. It's the one with an actual heat sink on the north bridge and the plastic socket lever. The BIOS is the latest from Tyan, 2.06 I believe. I'm tried it with both 1 and 2 1.2Ghz Palominos (AthlonMP). I'm using the NMB power supply. The memory is Crucial ECC-registered.. According to Crucial's web site the DIMMs I have are specifically for use with this board.

Attached I have an ATA-33 12GB hard drive, and two 18GB SCSI-160 IBM ultrastars attached to the onboard SCSI-160 A controller.

I've tested both with the onboard ATI video, and with an AGP GeForce3. I've also tested both with an without an SB-Live! sound card.

In every single possible combination of the above parts, I get severe data corruption when using the onboard IDE controller. Although it correctly detects my IDE hard drive it generates constant ATA timeout errors and slows the system to a crawl. This happens even when I dumb down the bios to use older PIO (non-DMA) modes.

Eventually I became so frustrated with it that I disabled the onboard IDE ports and stuck in a Promise ATA-100 controller. Bada-bing, the problem is solved. The the system screams *and* is quite stable. It's good enough for me this way, but I would have expected a 500$ motherboard to have worked out of the box.

The only problem I may have is keeping the CPU's cool. Even with thermosonic thermoengine heatsinks with good fans the CPU temp is still 50 deg C. Of course, part of that might be the 32 C room temperature...

Serverworks IDE chronically broken

[bbk]

The serverworks IDE chipset has, to put it lightly, some problems. It's DMA support is broken on some chipset revisions, so running it in anything greater than PIO mode 4 is a bad idea.

I'd buy a 3rd party IDE interface (3ware would be my first choice - lots of ports, modern drivers in kernel), or SCSI drives for the board. Tyan boards are great BTW - my old Dual PentiumPro Titan Pro ATX is running great after 6 years of use.

BBK

Just buy another board?

[duffbeer703]

There are alot of motherboard vendors...

Just ditch this piece of shit Tyan and buy something that works properly... MBs only cost like $100