Memtest and strange memory/bluescreens

[drwho9437]

I've been getting lots of blue screens, page file or IRQ typically. The system (Vista based) will then go into a loop of rebooting at times.

It appears this loop can be broken by a hard turn off of the computer, waiting a few second and turning everything back on. The reboot switch will not fix it. Thus I suspect a hardware related issue as you hear the soundcard discharge if you do the hard off, whereas on the reset though the post sequence it does not do that.

Now I checked my memory with Memtest+ 2.01 and when it was in one of these cycles of death (without the hard off), it managed 650 errors by subtest 4. All near 63.9 MB or 2044MB, it is duel channel and interleaved so I don't know really where those are on the two DIMMs.

Now I reboot with the hard off. Run Memtest+ no errors so far at all and I'm done the first pass.

So, I'm not sure if it really is the memory or if it is the motherboard I guess, or what do you think it might be. It worked very nicely for about 3 weeks (after building the system, 0 crashes then a few a day, it will work for hours at a time fine but decide not to work and crash crash crash).

Edit: Let me add the obligatory information

Motherboard Gugabyte Ga-P35-DS3L
Ram 2x Crucial DDR2 PC6400 CL4
CPU E2200
Stock cooling.
Antec Case 300 W
2 320GB Seagate 7200.10 HD, On-light DVD burner of some sort, 3.5 floppy
Chaintech Envy based sound
GeForce 6200 (I think) 256 MB
Vista SP1

 

[jaqie]

It's quite likely that it's the mobo, however there is no way to be sure short of dropping the parts into another PC. The main suspects here are the mobo (chipset) and the RAM. It's way less likely but possible it is the CPU or video.

As for the sound 'discharge' that is quite normal, and is no indication of something being wrong unless it has never done that before, even if it hasn't it could just be the state the sound card electronics are in when the system is in this state, and completely harmless and meaningless.

 

[drwho9437]

As for the sound 'discharge' that is quite normal, and is no indication of something being wrong unless it has never done that before, even if it hasn't it could just be the state the sound card electronics are in when the system is in this state, and completely harmless and meaningless.

I know this, I was simply trying to explain to others that a hard power off lets all the caps discharge so the machine state is different.

I'm not sure what the fault is. I hate unreproducible errors and this one defies simple reproduction. The time to fault and sequence are random. All I know for sure is that once it starts happening memtest will fail (2-3 verification). This could still be motherboard is the RAM controller is flaky. The time scale for a stable boot to crash is on the order of hours (happens over night pretty predictably)

My power rails look stable and fine on CPUID Hardware Monitor. The temp of everything is fine except perhaps the video card, but it has always run very hot.

I assume turbocache a kernal driver type of feature.

 

[jaqie]

Originally posted by: drwho9437
My power rails look stable and fine on CPUID Hardware Monitor. The temp of everything is fine except perhaps the video card, but it has always run very hot.

Rule #1 of computers: NEVER trust software monitoring of any type. They are wildly inaccurate as a rule (before calibrating them yourself) and should only be used after manual verification and calibration with real testing equipment.