Memtest and false positives

For topics which do not fit in other specific forums.
Piplup87
Posts: 11
Joined: Sun Mar 14, 2021 8:27 pm

Memtest and false positives

Unread post by Piplup87 »

Hi there.

Not sure if this is the right place for this question, happy to move it if there is a more appropriate category.

My setup that I got last month is running great with a 128 ram board, IOboard, and powered usb. Played for hours at a time on various things without issue. Except... the memtest is giving awful results (drops to 70mhz as soon as it can then the bottom red number starts sprinting upwards). I've re-seated it twice and cant see any issues visually, but I've never done soldering or electronics work so wouldnt really have the expertise to know what to look for.

Searching around I saw some posts on reddit discussing false negatives (a good memtest result but problems with cores). Are false positives a possibility with this test as well? I think I'll ask for an exchange for a replacement part to be proactive, but wasnt sure how much faith I should be putting in the memtest if that's the only sign of an issue? Thanks very much.
User avatar
aberu
Core Developer
Posts: 1170
Joined: Tue Jun 09, 2020 8:34 pm
Location: Longmont, CO
Has thanked: 244 times
Been thanked: 406 times
Contact:

Re: Memtest and false positives

Unread post by aberu »

If it's not passing the memtest, then it's only a matter of time before you notice a problem. Higher bandwidth requirement cores will stress it more, you might just not have used them yet. Have you used Neo Geo's core for instance? The test is just seeing how high a bandwidth your RAM can run for a certain period of time without errors. The numbers that are increasing are just it testing every single memory address. That's not going to give a false positive.
birdybro~
Piplup87
Posts: 11
Joined: Sun Mar 14, 2021 8:27 pm

Re: Memtest and false positives

Unread post by Piplup87 »

Neo Geo I've maybe put in 20 mins max before resetting into metal slug 1&2. Some slowdown with a lot of sprites (as expected though...)
I think you've answered the question though in that it's not possible for there to be a false positive, and even if I havent noticed it yet there will be a problem at some point.
Thanks, appreciate the response.
User avatar
aberu
Core Developer
Posts: 1170
Joined: Tue Jun 09, 2020 8:34 pm
Location: Longmont, CO
Has thanked: 244 times
Been thanked: 406 times
Contact:

Re: Memtest and false positives

Unread post by aberu »

It would be possible for a false positive to occur if something was wrong with memtest core in general, but we would have seen a lot of people having false positives I think if that was the case :)
birdybro~
Post Reply