Horribly unstable machine
Not directly Linux-related, but it's a Linux /box/. So please?
I just upgraded the mainboard in my old desktop after the previous board's cheap Taiwanese capacitors exploded; it's now an Athlon XP 1800+ running on a Biostar M7VIG Pro board (VIA chipset). After about an hour or so of uptime, the machine suddenly starts throwing segfaults left and right and dying. The BIOS temperature sensor showed the temperature of the processor at 140 degrees; that seems rather uncomfortably warm to me.
I have a couple questions for you all here:
a. I suspect I might not have centered the thermal tape on the heatsink on top of the processor die correctly. I need to pull the board out of the case and do that again, don't I. (ugh.) I have some Type 44 thermal compound handy if I need to... could a misaligned heat sink be the problem?
b. What should the CPU Frequency Selection jumper be set for? If it's open, it's 133MHz; if it's shorted, it's 100MHz. It's shorted by default; which should I use? (Overclocking is Bad.)
Thanks, all.
Update: Just because this entry keeps getting comments. It's been fixed. I used some Type 44 thermal compound instead of the thermal pad and the temperature dropped a lot. Maybe I was just centering it right this time or something. The heat wasn't causing the instability though; the problem was that I was using PC100 RAM when it was expecting PC133, which it didn't like even when the FSB was set to 100MHz (it *really* didn't like being at 133). So I bought some cheap DDR RAM (the board supports either kind; two slots for each type) of a known-good configuration and now it's been up for more than a week. Thanks particularly for the comment on the M7VIG Pro thermal sensor; I had suspected as much based on what it told me room temperature was.
I just upgraded the mainboard in my old desktop after the previous board's cheap Taiwanese capacitors exploded; it's now an Athlon XP 1800+ running on a Biostar M7VIG Pro board (VIA chipset). After about an hour or so of uptime, the machine suddenly starts throwing segfaults left and right and dying. The BIOS temperature sensor showed the temperature of the processor at 140 degrees; that seems rather uncomfortably warm to me.
I have a couple questions for you all here:
a. I suspect I might not have centered the thermal tape on the heatsink on top of the processor die correctly. I need to pull the board out of the case and do that again, don't I. (ugh.) I have some Type 44 thermal compound handy if I need to... could a misaligned heat sink be the problem?
b. What should the CPU Frequency Selection jumper be set for? If it's open, it's 133MHz; if it's shorted, it's 100MHz. It's shorted by default; which should I use? (Overclocking is Bad.)
Thanks, all.
Update: Just because this entry keeps getting comments. It's been fixed. I used some Type 44 thermal compound instead of the thermal pad and the temperature dropped a lot. Maybe I was just centering it right this time or something. The heat wasn't causing the instability though; the problem was that I was using PC100 RAM when it was expecting PC133, which it didn't like even when the FSB was set to 100MHz (it *really* didn't like being at 133). So I bought some cheap DDR RAM (the board supports either kind; two slots for each type) of a known-good configuration and now it's been up for more than a week. Thanks particularly for the comment on the M7VIG Pro thermal sensor; I had suspected as much based on what it told me room temperature was.
