Maximum PC

I'm running memtest86 v4.0a on a new build. I left it to run overnight. When I last checked it was on its 5th pass with no errors. When I check it again in the morning the screen looked like this:


Does this constitute and error?

Can anyone explain what I'm looking at?

Many thanks.

Can you shrink the picture? My system won't show it because of its size.

I liked it big, read it just fine, and now have to squint to see the same jibberish errors...but oh well. It's big and small now to make everyone happy.

You're going to get errors in Memtest. Errors are a fact of life with anything on this world, so why would desktop memory be exempt? This isn't ECC memory and this isn't a build that needs an endless Memtest loop to prove it's worthiness in a very healthy build. I just deleted a bunch of stuff to keep it short, but you're OK if it ran for just an hour. Even then, errors are expected.

Remember the P4 with it's high cycles and long line? If one hiccup happened it'd take a few nanoseconds to recover and go back through that long line again. Here, you're cycling a memory error to a point of failure. It may seem drastic...but what else did you see? The computer running through it all? Yup.

It's a blind test when you don't know what it's doing. It's a blind test because someone told you to do it and didn't explain what to expect from the readings...you just did it. Those that told you to do it probably heard it from someone else, who heard it from someone else, and really just don't know why this would be a test at all. Nothing against you though (you've just been assured your memory is good), but for those ass hats that blindly advise to "do memtest runs and see if it works" then leave you in the dust.

It's a stupid test and validates nothing on it's face. It's a great test for specific problems best left for experts with assistance. In this case, you took free advice from a random person that's just repeating stuff they read without actually dissolving the "why". Don't worry, it's not your fault for looking, but I'd take those free advice givers a second look next time.

They can ruin forums.

I'm going to have to disagree with Chumly on this one. Having memtest error out after any number of passes is not a good thing regardless of it being ECC memory or not desktop memory has gotten much better and should not error out at all. Memtest86+ is one of the standard burn in tests I run along with LinX to make sure that my system is stable. My current system ran 20 passes of LinX (considered the gold standard) without errors along with running memtest86+ with no errors. What you're looking at on the screen is garbage data that really won't tell you exactly what went wrong other than memtest86+ got some data thrown at it that it could not handle so it halted.

Now if you're overvolting/overclocking the memory you might want to bring it back down a bit and see if it'll run memtest86+ overnight again (although in truth 8 hours should be sufficient). If you're running it at stock or letting the BIOS automatically select the speed & voltage based on SPD data you might want to look up the specifications for you memory and make sure that the BIOS is clocking it correctly and sending the right amount of voltage. Failing that you will want to take out your RAM and start testing it one stick at a time. Run each stick for 8 hours and see if one has the same thing occur again or not. If none of the sticks do then you may have an issue with one of your slots, at that point even if you find the slot that's having the problem and are able to avoid it I'd probably RMA the motherboard as who knows if the other slots may develop the same issue.

Again, having memtest86+ error out is NOT something that's common and should be taken as an indicator that something is wrong with your system. Either the memory is clocked/volted too high, one or more of the RAM sticks is defective or something is wrong with one or more of the RAM slots.

Thanks Late Night,

I failed to say in my initial post that the screen capture I posted showed Memtest86 while it was STILL RUNNING. It wasn't frozen or stopped. It was still cycling through tests but as you see there was a lot of garbage on the screen. Don't know if that makes any difference in your assessment. For now I will assume that it was a genuine error and double check the ram setting and voltage (which I should have done before running the test).

That is odd and not something I have seen before with memtest86+. I'm going to check into that a bit and see what the typical cause of garbage data like that being displayed on the screen while the program is still running is. Normally if there is an error it should be displayed in the lower half as part of the log showing when the error occurred. It may still be an error or it could just be that the particular version of memtest86+ has some issues that have yet to be resolved or it has a compatibility issue with your motherboard. You may also want to try another memory testing tool to see if it shows any errors. Another good memory testing tool that I would recommend is HCi memtest. You might want to run it and see what happens.

Btw it's nsafreak, no offense taken .

If it passed 5 passes then it is most likely fine, memtest isn't the most stable program I have used. Half the time it doesn't even identify the processor or motherboard correctly. I run 2-4 passes and if it passes those the ram is fine.

Try comparing the MD5 hash of the ISO you have downloaded (and burned) with the hash they have on their website. If there was a glitch in the download or the burn, then it might spit out garbage like this. Also try reburning the disk at a slower speed, and telling your burning program to verify the disk.

Ok well apparently I'm an idiot on top of everything else. I purchased DDR3 1600 RAM when my Intel MB is rated for DDR3 1333/1066.
So my question is now, how do I know figure out how to manually set the timings given that I'm using 1600 rather than 1333? Or should I simply sell the 1600 on ebay and get some 1333 RAM?

Here is the link to the spec sheet for the memory I purchased:

http://www.kingston.com/dataSheets/KHX1 ... K2_8GX.pdf

nsafreak: sorry about that!
Danthrax66: I tend to think you're right. But I'm anal enough I like to see it pass a 12 hour test if possible.

Hmmmm that does change things a bit. Unless the memory is really finicky for some odd reason it should be fine running at below clock settings as long as it's getting enough voltage. Overvolting/overclocking memory is where you're more likely to run into issues than you are underclocking although undervolting can definitely cause issues as well. Make sure it's running at the highest speed that your motherboard will support and it's getting the right amount of voltage.

Nah you're fine I don't even run memtest unless I have issues. And that is just the official rating you can still run the memory at a faster speed.

So the latest version of memtest allow you to run the test in the current release (4.0) or opt to run legacy version 3.5. I ran 3.5 for 12 hours with no errors so I'm going to take dantraxx66's advice and move on with my life. Thank you everyone that took the time to help.