Windows crashes caused by bad RAM?

Ferris

Joe_Bloggs wrote:

@Mr Skint
Thankyou for your constructive suggestion and thanks to amit_1 for the concern and boldness of the proposed solution.

i have read the microsoft articles and they are not much use... they basically say check for new hardware and make sure its ok. as for reinstalling... see below...

Joe_Bloggs wrote:

1) Motherboard. These things have version numbers. The motherboard has a bios program that can affect its stability so quote it if you can.

i have the Asus K8Vb motherboard. my bios is not the very latest available, but i am loathe to flash it if i suspect any problems with my hardware... if the flashing fails due to bad memory, suddenly i'm left with a useless box... and in any case, could a bios become out of date if there have been no changes to the other hardware installed?

Joe_Bloggs wrote:

2) The processor range can be determined from the motherboard make but it would save time if the exact model was stated.

CPU Type AMD Athlon 64, 2200 MHz (11 x 200) 3200+

Joe_Bloggs wrote:

3) Temperature. I have experienced the same problems through over clocking which caused temperature related problems. Can you monitor the temperature of key components during normal operating conditions ? A few utilities are available for this purpose. Motherboard Monitor is one example.

i have used Everest and Sisoft Sandra, and the highest i have ever seen the CPU go is 45c, and thats when it was being battered by a torture test.

Joe_Bloggs wrote:

4)Memory. What memory are you using. What front side bus (FSB)? how many sticks of memory and in which slots. What are the memory timings and memory voltage ?

the memory came ready installed in the system. the FSB runs at 200MHz (according to everest). there are two sticks of memory (slot 1 and 2), running at 2.5-3-3-8 (i slowed this down from 2.5-3-3-6 in the BIOS, but it made no difference). not sure about memory voltage, but i think its 2.5.

Joe_Bloggs wrote:

PS what happened from December to the present Date ?

i'm not entirely sure. as i mentioned, it ran solidly from may to december - no BSOD, no crashes, nothing. the only problem was occasionally it wouldn't 'wake up' from standby/sleep mode. then from december suddenly it started crashing. i thought it might've been the Nokia PC software i installed, but removing it made no difference. over the past three months the crashes have become more and more frequent.

there is one other problem with the PC (started around the same time), if its any help. every so often, i am unable to start any programs in Windows XP- it just gives an error message saying "Procedure entry point not found" and mentions some kind of dll file - after a restart all is (usually) well.

i suspect reinstalling windows will sort this second problem out (maybe corrupt memory has corrupted some system files?) but i am reluctant to attempt a format and reinstall while there is a possible memory problem - i can just imagine regular BSODs in the middle of a reinstall *shudder*

on top of everything else, my wonderful shiny hologrammed genuine Windows XP Pro SP1 OEM CD has gone wrong, so I will have to locate a backup copy from someone somewhere (anybody help?) as i assume my OEM product key will only work with the OEM CD

thanks all for any help you can offer me,

Ferris.

Toxteth_OGrady

Try setting Vdimm in the bios to 2.7V and rerun Memtest

:cool:

TOG

djohn2002uk

If you are suspecting your memory, why not remove one stick and try that. If no improvement replace and remove the other and try again. Perhaps run MemTest each time. Unlikly that both have become faulty at the same time. If there's no improvement with either one removed, my guess is you're going to have to start looking elsewhere. Processor?

Ferris

thanks for all suggestions so far. does anyone have any idea why programs like memtest will find no errors for pass upon pass, and then suddenly find one or two? for example, i let memtest86+ run overnight (6 hours) last night, and after 3 hours it had found no errors, but an hour later it found two errors (on the modulo-20 test), and then found no more.

and how come one test finds errors but the others don't? surely if memory is faulty its faulty?

i am reluctant at this stage to mess with the hardware too much for fear of invalidating the guarantee. i just need to confirm that this really is a hardware issue, so that i can get the OEM to deal with it!

Joe_Bloggs

@Ferris you seem to know your stuff and appear technically aware of all pertinent matters. Computers can be educative in unexpected ways. Have you considered an fsb drop to 195 Mhz ?

You have a system that is on the margin of operating correctly according to memtest86. This test is the first hurdle that is easiest to leap. When you turn on the 'whistles and bells' in windows then other systems access memory directly. It a tough job to keep up with the demand. Prime95 is a fine example of a torture test and 3D games or 3D benchmarks are even worse on stressing system components.

I have never heard of 'Everest' the computer monitoring utility. Forgive my ignorance.

You have not mentioned the make and model of your PC. I hope you can get the technical support and replacement parts that you need .

Did you try a slightly higher ram voltage ? Often this can make the difference. However If your PSU is under strain due to failing capacitors on the motherboard or within your PSU then this advice would be little help.
J_B.