Memtest86+ 5.01 - reading the results

[Just]

Hello,

Those of us having already tested many times their RAM sticks with Memtest86+ may help me at understanding the results I got here (see screenshot).

​

Does it mean that :

• 
  one error appeared once during the fifth test of the first pass, and that is all
  Count : 1 at the right end on the last red line

or

• 
  one same error in Test #5 returned 4 times over 28 passes at the very same address
  Errors : 4 at the right end in the middle of the screen

I would rather conclude that the same error appeared four times at the same place (i.e. the RAM suffers of one hard error), otherwise I do not see what might be the meaning of the "Errors: 4" field. But I prefer to get a smarter feedback from Sysnative members. Understanding the meaning of the "Count: 1" column would be great as well, as I would have expected to read here "4" instead of "1"; as this is not the case, what does that field count for ?

Thank you.

 

[Digerati]

You are right, it is confusing and frankly, I am not sure what that means, except that your RAM is bad. My guess is you have one bad memory location that caused 4 errors.

Nevertheless, you are using an old outdated program that has not been updated in years. I cannot even find any information on that version any more. It does not help there similar programs called MemTest, MemTest86+ and other like sounding programs.

I recommend you download and run the latest version of MemTest86 from PassMark. This program is under constant development with regular updates (the latest just Feb 27, 2017) to ensure it supports the latest RAM technologies.

And again, even 1 error is 1 too many.

You did not say what version of Windows you have but later versions of Windows include an effective Windows Memory Diagnostic tool too.

 

[Just]

Thank you for your answer.

One day later, memtest86+ showed up 6 more errors for a total of 10, after 44 passes, but not a single more failing address. It goes in the same direction as your conclusion.

​

I then tested the failing stick alone. Again it returned an error from the very same address, after the 9th pass. That second test confirms that :

• the error got caught more than once at the same address, which makes sense to see an "Errors:" counter rolling up and not being sticked on "1" as the "Count" column was ;
• the memtest86+ "Errors:" counter located at the far right in the middle of the screen is the one to look at,
  not the "Count" column;
• up to 9 passes may have actually been required to identify that hard persistant-but-flickering error, which a shorter test would have kept unnoticed. Keep memtest running longer is good !

​

Finally, what does the "Count" column shows up ? Let's the question stay open for the next time. :lol:

 

[Digerati]

Again, I fail to see why you continue to use an outdated program when up to date options are available. Maybe the count question will never be answered because the program is no longer under development.

 

[satrow]

Once you complete run 8, stop. The chances of spurious results that wouldn't affect the system in normal, or extended uptime, usage begin to increase from there on.

Using hardware testing software from 27/09/2013 on hardware that's even older isn't an issue, the same software is still very good on DDR3 -equipped systems (at least), and, afaik, is still the main memory testing software suggested by most BSOD forums/experts. Windows own memory testing routine is very basic in comparison, not worth wasting your time on.

 

[Digerati]

I have no doubt the old software will work with older RAM, and if it reports problems, the problems are real. As I noted above, what was certain in my mind was that the RAM was bad.

But old software can still have bugs and newly discovered bugs and the fact it is no longer under development means any bugs are probably not being addressed. A program that is still in current development is likely to kept current.

I do, however, disagree with your assessment that Windows Diagnostics is not worth the time. While no software based RAM testing program is 100% conclusive (that takes specialized and expensive test equipment), Windows Memory Diagnostics uses the same testing protocols as other testers. It is basic in the way most integrated Windows tools are basic - that is, no fancy UI. It has been around since Vista. If it was not worth using, MS would have dumped it. Like other RAM testing software, if Windows Memory Diagnostics reports the RAM is bad, it is bad.

That said, it would not be my first choice. MemTest86 from Passmark is - which BTW, comes from the same roots as MemTest and MemTest86+.

 

[satrow]

Bill, maybe you should start at 'home' and convince everyone that the BSOD posting instructions need a rewrite.

 

[Digerati]

There are a lot of instructions that have been around for years that need updating. I'll leave keeping BSOD posting instructions up to the BSOD analysts. BTW, being a BSOD expert does not make one a hardware expert anymore than knowing a bit about hardware makes me a Windows expert. And of course, a BSOD does not automatically imply defective RAM any more than defective RAM automatically cause BSODs.

But none of that changes the facts which remain the same. One program is out of date and no longer under development. The other, which is built on the same base testing methodologies, is kept current with the latest technologies to support the latest hardware advancements - as seen here.