Crashes driving me nuts

[Cornbuds]

I built my current main desktop about 5 months ago (specs in sig) and I get random but regular crashes. I'm getting both application crashes to desktop (BF2 is the most frequent cause here), but I also get regular, but less frequent BSODs - again BF2 causing most BSODs, but other apps as well. I know BF2 is not exactly the best predictor of system stability, it's a relatively frequent problem for a lot of users, but something is definitely not right.

I'm running everything at stock timings/clocks because I'm trying to find out what's causing my problems. My RAM is running 2-3-2-5 @ 2T (I get the problems at 1T and 2T), which is SPD according to OCZ. My RAM part# is OCZ4002048ELDCPE-K.

The most frequent BSOD is PFN_LIST_CORRUPT, which is generally associated with bad RAM if hardware is the culprit. I've run memtest86 for 72 hours straight with no errors reported. I also ran Prime95 for 72 hours with no warnings or errors reported, although Prime95 did crash my OS another time I ran it, but I can't remember which torture test was running.

I'd sure appreciate any advice anyone can offer, I'd hate to RMA my RAM if it's not definitely defective.

Thanks

 

[Flip-Mode]

What PSU are you running?

 

[avalanche83]

Do you have the cpu overclocked? If so then you may have to back that off a little. I can't see the ram being what is causing the BSOD's.

 

[rseven]

If you failed the Prime blend test then that could point toward the ram. However, a faulty PSU could be behind this as well.What kind of PSU are you using?

 

[Cornbuds]

Nothing is overclocked right now.

My PSU is an Enermax 431W unit - part# EG465P-VE, I've had for about 2 years. All my temps are very good as well (case, CPU, GPU).

Thanks for the responses.

 

[avalanche83]

Is all of you settings right in BIOS? What is you LDT and voltages at? Have you tried increasing the vdimm?

 

[bchur83]

I think that PSU might be a bit on the weak side. It probably has a high 5V rated rail, and not so much 12V. I would try a different PSU.

EDIT: I looked up your PSU, and it shows as being a 460W with 33A@12V. That looks like it should work.

 

[Cornbuds]

My PSU is definitely a 431W unit. Enermax is kinda retarded when it comes to their part#'s, it seems they recycle them when newer units are released. Mine is not the blue "noisetaker" series, mine was out before the noisetakers were available. I'm reading the wattage and part# right off the PSU. This is my unit. When I Googled my part#, I do see the newer blue coloered units returned also. My PSU has a 44A 5V rail and 20A 12V rail. You're right bchur83, my 5V rail is much higher. Is 20 Amps not enough for my drives? I was under the impression it was.

I've triple-checked all settings in my BIOS. FSB is running stock 200MHz. Vcore at stock 1.400V, vDimm at SPD of 2.60V. I did try incremetally upping the vCore and vDimm (separately) and still get the same results.

Thanks again.

 

[bchur83]

Not is it so much the drives draw too much power, but more so the entire system. The Vcore is derived from the 12V line, you video card runs off the 12V and so do most of the other components. 20A is probably not enough. What voltages does your PSU give? Check in the BIOS, or better yet, use a Multimeter. If they are lower than the rated voltages, then the PSU may be the cause. Also, check with a MM when the system is under load. That would be the best determining factor of a weak PSU.

 

[Cornbuds]

I came across this article today after your input about the PSU possibly being the culprit. Seems pretty well done yet indicates the vast majority of systems never exceed 200W total. It indicates less than 10A pulled off the 12v line.

Any thoughts?

 

[rseven]

Interesting article. I didn't have any idea what PSU you had, and certainly a weak one could cause such a problem. However, the Enermax has more than enough juice to power your system. This leads me to to lean toward it being your ram. I know you don't agree, but it is obviously something causing it. What I would try is to go back to 1t command rate, yes, you heard me right, but I would loosen the timings to 2-3-3-6. The faster command rate will easily make up for the slight performance hit, and the overall timings may be more in sync with your system. Sometimes ram will not perform at it's fastest timings in dual channel. There should be no reason it will not run at 1t, but the other timings could be an issue. If this is not the case then I would start pointing fingers at the MSI mobo, which have never had a Quality build reputation. Good luck with it.

 

[Cornbuds]

I would start pointing fingers at the MSI mobo, which have never had a Quality build reputation.

/hijack
LOL, "quality" is such a subjective descriptor. This MSI board has been a royal PITA, buggy as hell from the get go. BIOS flash 1.9 wouldn't even POST, fortunately I had the old BIOS (1.4, I think) and reflashed successfully. Went back to MSI's site and BIOS 1.9 was mysteriously gone. PS/2 port for mouse won't work with a USB keyboard, and on and on. I believe this will be my last MSI board, that's for sure. Only reason I got it was for 8 SATA ports and dual onborad RAID controllers.
/end hijack

So is it relatively safe to say if memtest86 runs for 3 days straight (I can't remember how many passes it made), and it reports 0 errors, my RAM is most likely good?

I'm gonna rerun all 3 Prime95 tests independently and see if I still crash. I'll also try your suggestions rseven.

I appreciate everyone's input.

 

[rseven]

Unfortunately, running memtest86 for a week does not mean that it will run stable in Windows. Windows just puts much more stress on the ram and it's timings. The Prime blend test is really much better at determining the rams stability in Windows. Another possible issue is the ram may not be getting the voltage it requires. Many boards undervolt slightly. I know my Neo2 does. Add this to a normal dip under stress and you could have a ram failure. I would start with the Prime blend and go from there.

I'm not surprised to here of your MSI woes. I have been preaching that theme here for the past year. Good luck with it.

 

[Brando]

This reminds me of a problem with my stepfathers computer awhile back. I turned out the heatsink on the video card had come loose from the pushpin that held it on. He would play a game for awhile then crash to desktop when the heat became too much to handle.

 

[dfonda]

Does that Ram want to run at a higher FSB maybe?

 

[T'wolves]

I'ld still get that multimeter out while your running memtest and prime. check the 12v especially, the 5 also, and a 3.3 if you got access without moving stuff in a running computer. If your running prime and or memtest and your gettin 11.6 or 11.5 on that 12v rail, you got problems.

 

[Cornbuds]

Does that Ram want to run at a higher FSB maybe?

Not much, it won't go past 215Mhz, even at veeeeery relaxed timings, which was kinda disappointing. The mobo runs 250FSB/166MHz/4x multi/2-3-2-5 1T as stable as with everything at stock settings, which includes the crashs from time to time. FSB maxed out at 275MHz/133MHz/3x multi/2-3-2-5 1T. (Obviously with more juice on everything)

I'll grab my multimeter and check the lines themselves, but my instincts tell me it's not my PSU. I'll post back with the results.

Thanks again for the input.