• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Memtest testing all 4GB?

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Category 5

Member
Joined
Mar 23, 2005
Running memtest+ now and noticed that it is on test 7 and it says testing 108K-3072M 4096M.

Before I thought I was going crazy when i saw it was 18% on test 4, and came back a while later and it was on 4% of test 4 and still on the same run. I am guessing what is happening is that it is running each test for the first 3Gigs and then running it again for the last Gig separately. Rather than sit there and stare at the screen though I have my laptop hooked up to the same monitor so I am switching back to check for errors every now and then.

Can someone shed light on this behavior? I can't imagine it wouldn't test more than 3Gigs since memtest would then be useless for modern systems. I just want to make sure that when my test is up (24 hours) I can be sure all 4 1Gig sticks have been tested. I need to rely on this PC!
 
After further examination, it looks like memtest IS first testing three gigs, and then the fourth afterwards. It is also misreporting some things. The first run of each test states (as above) that 108K - 3072M are being tested. Thing is, the progress bar counts incorrectly and instead of going up to 100% it goes up to 20% or 25% depending on the test. then, it tests the last gig, at which time the progress meter works correctly (goes to 100%), but the description then says testing 4096M - 5120M. Of course the system only has 4096M, but since no errors are being thrown it is actually testing ram cells, and they are passing.
 
I'll answer myself again. After some reading I found out that aside from the test % meter reading incorrectly when testing 3 or more GB of RAM, the second tier of the test (testing the last GB of my 4GB system). It reads that it is testing the 5th (non existant GB) GB because that memory is shifted to those addresses. It is done because the MB reserves the adresses in between for the PCI bus and system operation.

Just in case someone else had a question about this I thought I'd post my findings. The only incorrect behavior is the meter, but according to concensus the test is still working just the meter is reading incorrectly.
 
Back