is this a hardware issue?

For some background on my computer, please refer the following threads on this forum:
http://www.hardwareforums.com/strange-problem-my-computer-9640/
http://www.hardwareforums.com/motherboard-amd-xp1800-firewire-9617/

Ok, now on to the problem: I went ahead and purchased A7V400 motherboard,
and for some reason my OS doesnt boot. Fair enough?, the hardware has
changed and I need to re-install WinXP OS. When I boot with my WinXP CD
I get this error:

ntkrnlmp.exe could not be loaded
Error Code 7

The microsoft KB article http://support.microsoft.com/?id=812580
suggests me to press F7 to skip ACPI detection. I tried this 10 times but I still get this error.

The link here http://www.computing.net/windows2000/wwwboard/forum/31272.html
suggests many things: many of them are about RAM. I have 512Mb Hynix
DDR333 RAM module. I tried re-inserting the module, putting it in the other
slot. But this did not help.

I am confused as to what I should be trying out next. How do I find out if
RAM module is faulty? I have actually tried with two RAM modules - 512Mb
Hynix DDR333 and I dont know what brand but it is 256Mb DDD266, and the
result is same above error. Both these modules have been working very well
with my earlier motherboard, although only individually.


Based on the messages in the computing.net forum, it looks like the problem
is resolvable. I am posting this message here to see if this indeed is a
hardware issue.

Thanks for the help,
--f

 

Try running Memtest86+ on that machine overnight.

 

I am currently running memtest 1.60 that I found on a Knoppix CD. Although it is still running I cant wait to post this and ask: On my 256Mb module it is currently showing 16384 errors, does it mean this module is not usable anymore and deserves to be thrown away?

Is it correct to say that unless the number of errors memtest reports is exactly zero, the RAM module is sure to cause random problems like the one I am facing?

 

I think you can assume that.
Memory is funny like that (what a Fred Flintstone way of putting it...LOL).

Even if it showed just one error after running for hours it is likely to cause problems at some stage... otherwise it would have not returned an error, you wif me?

I would put it aside and start again. Start without any memory and see how the computer POSTs. Plug the 512 module in slot 1 and run memtest. See what shakes. If the module is good try and boot... see how it POSTs. It could be that another peice of hardware is failing your system. The only way of finding out is by trying to POST using process of elimination...

 

The motherboard is new, so although possible it is less likely that something is wrong on the motherboard. the CPU and the memory modules are old.

The 512 module is giving far less errors, so far it is 293 errors. But I get the same error that I mentioned in my first comment with this module too :-(

 

I will strengthen AT's point made in several threads:

If your RAM returns even just one error, it is a cause for concern.
Let alone hundreds of them.

There is hardly any point in running a system which is relying on faulty RAM.

My suggestion to you is - if you know that one part is problematic, chuck it away and start with a new one. Otherwise you are tripping yourself with your own feet, or cutting the branch you are sitting on or...... you get the point.

Get a new Module and try it. If it does not help, run memtest again. If memtest returns errors again take it back. You have to acertain what it is that is hindering your system. Unless you have a perfect RAM module, or if you can prove that the one you have can perfectly serve another rig, you will never know whether it is indeed the cause to all your problems. You will just keep going in circles.

 

Steer clear of generic name brands. They may offer 'too cheap I can't lose' value, but they will probably cause major compatibility issues. They don't last as long as branded RAM, and don't offer the same kind of performance.

General brands I recommend are Kingston, Samsung and NEC

 

That RAM isn't necessarily bad, although it certainly may be. There are other possiblities presented by your scan results indicate that it might also be incompatible ram, incorrect RAM timings/clocks in your BIOS, or even a bad northbridge. As Sabashuali pointed out, the best way to troubleshoot from here would be to take some known working RAM (preferrably of a trusted brand name) and do another memtest in the system. Conversely, you could test out the RAM which returned so many errors in a known working system and see if you can reproduce the results.