Annoying lockups...

[GHoosdum]

My secondary system locks up after anywhere between 6 hours and 17 hours of constant running. I have no firm idea why it does this.

Current setup:
XP2100+ Chip, running at 2.25 GHz, 1.675-1.7 VCore, maxes out at around 43C. FSB 166, multi 13.5
2x256 PC2100 (Crucial Micron) DDR Running at 133MHz (total stock settings) dual channel mode
60GB Maxtor HDD
Gainward GeForce4 Ti4200 8X Golden Sample - running stock speeds
Asus A7N8X Deluxe latest BIOS revision - has run 166 FSB stable (weeks constant uptime) before without any issues.
WinXP Pro SP1

I'm having a bear of a time tracking this one down... I've lowered the processor clock with the same lockup result. I turned off the overclock of the Gainward card, and the problem went away for a few days, then it came back. I can't rule out the possiblity that the few days uptime was a fluke.

I used to think it locked up when the processor reached 46C, because that was the temp it read when locked up previously, but now that I lowered the VCore, it only gets to 43C max. It locked up today at 41C. I'm getting pretty angry at it now.

All this time, it's folding away Gromacs with -forceasm and -advmethods without a problem, except for the lockups. Zero errors in folding.

[a2jfreak]

No reason to run your FSB @ 166 when the RAM is at 133 since your CPU's multi is unlocked by the AXN8X. Run your FSB/RAM in synchronous mode and just up your multiplier. Also, your PSU could be failing due to insufficient cooling or any number of reasons. Also, have you cleaned the heatsink lately to make sure the air is getting to it rather than being blocked by dust? You might need to up the VCore another notch, or you just might be running the CPU a few MHz higher than it's capable.

Also, make sure you have the most recent BIOS.

[GHoosdum]

The reason to run my FSB @ 166 is that it can... running Async is fine so far, with the dual channel DDR there's no lack of memory bandwidth. I can't up the multi above 14 in this BIOS anyway, so there's no way I can reach this speed at 133 FSB.

The PSU gives fairly constant voltage, a bit above the setting in BIOS but within tolerance levels. Is there anything else I need to worry about with the PSU?

Heatsink is clean as a whistle, and the temps are (obviously) staying low.

I get these lockups at ANY VCore setting, from stock to 1.775V, and at every clock setting I've tried on the processor. It did it with the XP2500+ processor at stock speed, too.

BIOS is most recent version, 1006.

I still have no clue what the culprit is, but I highly doubt it's my overclock.

[muddocktor]

Try running the client without using the -forceasm switch for a few days so that the client uses 3DNow! instead of SSE optimizations and see if your lockup problems go away. I have an XP2400 that will not even do 1 Gromacs WU without locking up when using SSE but will fold for weeks on end when using 3DNow! optimizations. This proc used to be able to run some Gromacs with SSE but it seems like it's lockups got worse over time. It will lock up when folding using SSE even at stock speeds. I just accepted the fact that it doesn't like to fold using SSE.

[GHoosdum]

Like I said, it did the lockups with another processor also... wouldn't that rule out the SSE? Unless BOTH the 2100 and 2500 I have can't run with SSE!?!? :banghead:

[muddocktor]

Both Tbred B's and Bartons can have problems with using SSE when folding Gromacs. Just give it a whirl for a few days and see if the lockups go away. I know that it will decrease your folding by 10-15%, but you really need to find out if that is your problem.