my Memtest86 results - 4 errors/9 hours - bad or normal?

DarthWindu · 8th September 2006, 8:55 AM

heyas

i'm wanting to get some feedback on my memtest86 results.

I have been having random BSODs recently. They have all occurred since a recent format and upgrade of my primary hard drive (to a 150gig raptor). otherwise my hardware has remained static for the past 36 months, and my system has, historically, always run flawlessly. i have had probably 10 BSODs in the last month - one was when i was transferring files to an ipod, one was soon after loading in to WoW, last night it was while in the middle of watching a divx. no real pattern that i can discern.

athlon64 3400+
x800xt
1gig cheapo kingston ram, ddr400, 200mhz, cas 2.5/4/4/6 (or is it kingmax? hmm can't remember now..posting from work)
asus motherboard, again can't recall the model number.
soundblaster live

latest drivers for everything, i believe.

i ran Memtest86 overnight last night, to see if my RAM might be giving up the ghost, since RAM seems to get the blame for most BSODs as far as i can tell.

almost 9 hours of testing, 4 errors came up:

Tst/Pass/Failing Address/Good/Bad/Err-bits/Count

5/3/0000824a6e8 130.6MB/efffffff/ffffffff/10000000/1
5/3/0000924a6c8 146.6MB/efffffff/ffffffff/10000000/1
5/3/0002890a4a8 649.6MB/efffffff/ffffffff/10000000/1
5/3/0002990a488 665.6MB/efffffff/ffffffff/10000000/1

does this indicate a problem with my ram? or does faulty RAM give off a million errors if it is a real problem?

not sure what i want to hear...that i have to fork out cash for new ram, or that its not the ram but i have to track down some obscure software problem >< and of course, i don't want to replace the ram unless that is definitely the problem...

thanks for any advice...
cheers
dw

Mad 4 Air · 8th September 2006, 9:44 AM

Any memory errors are bad. Try bit more voltage.

Lem0n · 8th September 2006, 9:45 AM

I'd suspect that even though you only have 4 errors after 9 hours of testing, the RAM will be giving you problems in windows. I have read and experienced RAM passing memtest for hours and hours, but once you load windows and try and stress it there, it fails.

Memtest will only give you an indication of how well you RAM will perform whe overclocking. Even though you're not overclocking, because it is failing at the stock frequency, it is faulty.

It won't be that bad if the RAM is faulty if it is Kingston, as they offer lifetime warranty, however that could be limited to what they think the life of the RAM is. You'll want to ring up Kingston and have a talk to them about it.

Goodluck.

DarthWindu · 8th September 2006, 9:53 AM

thanks for the replies.

i have some more potentially relevant info that i didn't think of including in original post, after bouncing off a suggestion someone else gave me...

and this takes it off the RAM track, so forgive the self-thread-hijack

my case is a zalman TNN500A passive case - passive 300watt power supply. adding the new hard drive (i still have my 200gig seagate drive plugged in as a data drive) may be straining the PSU too much...can't recall what wattage an x800xt draws, but i suppose it is a fair whack...

perhaps i should do some psu voltage monitoring...see if it's struggling at times...any suggestion as to how to do that best? any particular software that is handy for that?

re: more voltage for the ram - i am a tiny bit reluctant, given that i am and always have been running everything bog-standard, no overclocking, no fancy performance hax, or anything like that. and since it has always been running fine, not sure whether that may exacerbate the problem? would RAM develop over a few years, a need for more voltage?

unfortunately (or fortunately?) the BSODing isn't frequent enough to be able to identify what may or may not work to address the issue, other than noticing that i haven't bsod'ed for a month or what have you...

Lem0n · 8th September 2006, 10:09 AM

Well, you could be right about the Raptor overstressing the PSU as that is the only change that you've made since the problems started. Have you tried running the system without your Raptor?

Heres a link to a guide on how to test your PSU's rails --> http://forums.extremeoverclocking.co...d.php?t=137886
You'll need a multimeter to do this, but it is the most accurate way.

Mad 4 Air · 8th September 2006, 10:19 AM

Someone who actually runs one of those cases

. That is a whole new topic in itself.

The slight upping of ram voltage max fix your issue. I wouldn't consider .1v a performance hack

. Definitely try and get your hands on a multi meter and test the rails out.

I am i correct in assuming you can't mount a standard atx in those zalman contraptions?

DarthWindu · 8th September 2006, 10:35 AM

Quote:

Originally Posted by Lem0n

Well, you could be right about the Raptor overstressing the PSU as that is the only change that you've made since the problems started. Have you tried running the system without your Raptor?

Heres a link to a guide on how to test your PSU's rails --> http://forums.extremeoverclocking.co...d.php?t=137886
You'll need a multimeter to do this, but it is the most accurate way.

cheers, i'll see how i go. multimeters are not something i am familiar with, but what's the worst that can happ- o wait :O lol

Quote:

Originally Posted by Mad 4 Air

Someone who actually runs one of those cases

. That is a whole new topic in itself.

The slight upping of ram voltage max fix your issue. I wouldn't consider .1v a performance hack

. Definitely try and get your hands on a multi meter and test the rails out.

I am i correct in assuming you can't mount a standard atx in those zalman contraptions?

TNN rocks

i looove having a 100% silent rig (well, apart from the hard drives...which are not quite as quiet as my old Barracuda IV...but not enough to really hear...and the 92mm low speed panaflow that i run across the x800 during summer...which i started doing after i hit 86 degrees playing Far Cry on a heatwave day a couple years back lol)

wouldn't be able to mount a standard psu inside the case, but i'm not averse to having bits and pieces hanging all over the place if it became necessary...hardly ideal though.

i'll see what i can do about a multimeter, and see how i go. if i never post on ocau again, then please arrange for the burial of my charred remains

i'm almost tempted just to buy some new ram and see what happens, i could live with having 2gig ram anyway...since RAM isn't rediculously expensive at the moment...but yeah, i shall try to rule out the PSU before doing that.

thanks again everyone

cvidler · 8th September 2006, 10:44 AM

It's been ten minutes, you dead yet?

RAM and bad/insufficient power are the most common reasons for hard to troubleshoot problems.

Using the multimeter isn't hard, and unless you're silly enough to go poking around INSIDE the PSU, there's nothing there that'll kill YOU.

Mad 4 Air · 8th September 2006, 11:13 AM

Digital multi meters aren't hard to use. I know jack shit and i use one. One time i got lazy and tried to read 12v reading from a 3pin connector. Hand slipped and made a blue spark as i shorted 2 pins together. No harm done though

.

8th September 2006, 9:45 AM	#3
Lem0n Member Join Date: Feb 2005 Location: Perth, SOR Posts: 924	I'd suspect that even though you only have 4 errors after 9 hours of testing, the RAM will be giving you problems in windows. I have read and experienced RAM passing memtest for hours and hours, but once you load windows and try and stress it there, it fails. Memtest will only give you an indication of how well you RAM will perform whe overclocking. Even though you're not overclocking, because it is failing at the stock frequency, it is faulty. It won't be that bad if the RAM is faulty if it is Kingston, as they offer lifetime warranty, however that could be limited to what they think the life of the RAM is. You'll want to ring up Kingston and have a talk to them about it. Goodluck. __________________ Good Traders: spankyofoz, dug777, XST, Ellroy80, garbs, Wobbbz, Kayl, m4sopmod, rowen, dragokahn

8th September 2006, 10:09 AM	#5
Lem0n Member Join Date: Feb 2005 Location: Perth, SOR Posts: 924	Well, you could be right about the Raptor overstressing the PSU as that is the only change that you've made since the problems started. Have you tried running the system without your Raptor? Heres a link to a guide on how to test your PSU's rails --> http://forums.extremeoverclocking.co...d.php?t=137886 You'll need a multimeter to do this, but it is the most accurate way. __________________ Good Traders: spankyofoz, dug777, XST, Ellroy80, garbs, Wobbbz, Kayl, m4sopmod, rowen, dragokahn

8th September 2006, 10:44 AM	#8
cvidler Member Join Date: Jun 2001 Location: Canberra Posts: 7,206	It's been ten minutes, you dead yet? RAM and bad/insufficient power are the most common reasons for hard to troubleshoot problems. Using the multimeter isn't hard, and unless you're silly enough to go poking around INSIDE the PSU, there's nothing there that'll kill YOU. __________________ We might eviscerate your arguments, but we won't hurt you. Honest! - Lucifers Mentor ⠠⠵ [#]