Rindis.com

A few days ago, I was gifted/donated a computer. It would have been thrown out otherwise, and featured a better processor than I have, so I was happy to take it.

Examination waited for the weekend. Physical work waited for Sunday, as I had to get a longer-shaft phillips as the heat sink was screwed into place, and I needed to thread through the fan blades, and on one side needed clearance for an adjacent support bar of the case.

Mostly it’s a wash. I haven’t been able to determine the type of video card it has (has an nVidia logo, but none of the various serial numbers or part numbers have turned up anything in a search). The board is direct from Intel, and should be pretty solid, but it has no SATA ports that I can observe, and only 1 GB of RAM on board (my system has two). However, my current processor is an early 2 GHz Pentium 4 (second generation, the first Socket 478 chips), and this had a late-model 3.4 GHz Pentium 4. As my motherboard is comparatively recent (bought new when I built Haruhi a couple years ago), it is capable handling the chip (which is still on the same socket).

Or, at least, its supposed to be able to take it.

The swap out of chips between the two systems went very well. The only problem was that when I tried swapping out the graphics card, to identify the new one, I got a bunch of errors. I figured it was just a problem with that, deferred it to a later date, and started testing the new chip.

It’s surprisingly good. WoW has never been great on Haruhi, and as Blizzard has been adding more flash to combats and upgrading the graphics engine, its been getting slower. After getting the boost to 2GB of RAM, the primary culprit was the aging GeForce 6200. But with the new processor, WoW was looking pretty good. I still need to try out some of the places that were really causing problems, but some initial tests in places that had lots of NPCs around (the practicing swordsmen in Shattrath have been the bane of this computer) looked really good.

For about 5 hours. Then WoW crashed.

I’ve had crashes like this one before. In fact, I had them about 3-4 times a week at one point. But WoW crashed again a few minutes later, and the system rebooted.

BSOD on reboot. Screen full of errors on boot to Safe Mode (failed). BSOD on boot to the Win2000 console straight off the CD.

At that point I knew I was in big trouble. I tried all sorts things: disconnecting the hard drives, swapping out RAM…. BIOS reported the chip was nice and cool. The original error is pretty rare. The only references I could find to it were on sites in Japanese or Czech. Other errors (it started cycling through a number of BSOD errors) were clustering around the primary buses, and made me worried that I might have actually blown my board.

One thought was that I might need a BIOS update for the board to handle the new chip. However, MSI has put all their updates under a new updating system that means you don’t have to go through the entire process of flashing the BIOS through a floppy (which is exactly what I wanted to do), and it does a nice job of auto-detecting your board and current BIOS version. But you have to be able to boot the system for that, and they don’t let you at any of the support materials for other equipment….

So, finally I swapped the chips back. Maybe Haruhi would work again. If not, I could get the other box running and transfer my data hard drive over to it.

Haruhi booted.

This was a relief, if still annoying. Checking with MSI showed that I had the latest BIOS, but my chipset drivers were old. Since that would impact the types of things going on here, I updated them, and pulled Haruhi apart for one final try.

So far, so good. Considering the amount of time it took for the original collapse, I’m only cautiously optimistic, but with any luck, I’ve got it beat. I’m a little concerned about if I ever need rebuild the install from scratch…. And the plan is to build the new machine (with the old chip) into a backup machine, and then I’ll get rid of Utena, which has been sitting in a corner in case anything bad happens to one of the main machines.