Best RAM/memory test

[BTRY B 529th FA BN]

What's the latest best memory testing app? Currently using Memtest HCI. For some reason it found errors while Memtest 86 + wouldn't.

 

[Blain]

Try these...
DocMemory
Windows Memory Diagnostic

 

[Deders]

I second Windows Memory Tester, it too found things that Memtest couldn't.

 

[john3850]

Memtest HCI is very usefull for L3 cpu errors.

 

[BonzaiDuck]

Memtest HCI is very usefull for L3 cpu errors.

We crossed paths during the Sandy Bridge days. What do you mean "L3 cpu. . . ?" CPU's with L3 cache?

All these memory tests take time, and the extra time is linear with the size of the configuration. 16GB took four + days when I ran it!

I'd been using HCI Memtest-64. It seemed highly recommended when I first started using it on this machine. You can run it within Windows, or you can run the self-booting version from CD.

The earlier Memtest programs didn't seem to keep up with newer processors (hence, they might test the memory but couldn't identify the processor in the information posted on the screen while running the test.

There's always going to be something better. If I build my "E" system next year, I will no doubt update my stress-test and memory-testing options to what is "best advised."

 

[bryanl]

Have any websites published comparative reviews of them?

The newest I could find is this, with a publication date of Dec 2001:

http://www.realworldtech.com/memory-diagnostics-2/

 

[Blain]

2014 reviews

 

[BonzaiDuck]

2014 reviews

That's an interesting review.

I'm wondering why they didn't include HCI Memtest-64. I'm also surprised that Memtest86 and "+" are still in the running or still being used.

At one time, I had an earlier version of Windows Memory Diagnostic, probably using it here or there.

One always has to wonder why -- or whether -- a reviewer has made a comprehensive comparison, or simply chosen items to review based on some other logic.

 

[Deders]

It's not so much a review, more a "here are some options". No in depth testing has been done to find out which programs are likely to pick up on which errors etc.

 

[Batmeat]

HCI Memtest-64 <-------- This
if your machine is running a 32bit OS, then Memtest 86.

 

[BonzaiDuck]

HCI Memtest-64 <-------- This
if your machine is running a 32bit OS, then Memtest 86.

Remember "Treasure of the Sierra Madre?"

32bit?!! 32bit?!! Whadda we need stinkeen 32bit for?! We don' need no stinkeen 32bit!!

 

[ronitbajaj]

Memtest HCI and DocMemory are the best memory testing app. It's very useful.

 

[BTRY B 529th FA BN]

Memtest HCI and DocMemory are the best memory testing app. It's very useful.

Is DocMemory a tester on a PCI card? Or is there software like Memtest HCI ?

 

[BonzaiDuck]

It's not so much a review, more a "here are some options". No in depth testing has been done to find out which programs are likely to pick up on which errors etc.

And that's important. Errors would be more sparse and few between the longer the test program ran at a certain stress level. If the stress levels are weak, I'd think the greater likelihood of an insufficient test.

 

[bryanl]

Memtest HCI and DocMemory are the best memory testing app.

Review?

My old link indicated neither being as good as Memtest86 or RST, with DocMemory being particularly bad, and a few years ago the makers of DocMemory surprised me by admitting Memtest86 was better.

Outdated, but still the only memory diagnostic reviews available:

http://www.realworldtech.com/memory-diagnostics-1/2/

http://www.realworldtech.com/memory-diagnostics-2/2/


Memtest HCI is referred to as just MemTest.

BTRY B 529th FA BN: DocMemory is available in two forms, a card (I see only PCI) and software: http://www.simmtester.com

 

[fire400]

something so odd i don't trust about memory tests, even if the hardware is bad with a very specific defect or flaw, no memory test in short term can pick it up, so sometimes i just use common sense. windows xp worked fine, but win8.1 was gui flawed, and i just swapped out for good samsung RAM.

on my thinkpad laptop, i was getting sketchy lines across the gui. i'm like, i just tested this in different OS environments and even a boot CD, with different kinds of memory burners.

i didn't waste my time, trying to uninstall graphics drivers or updating drivers, i just swapped the RAM out for a diff. dual matching set of samsung's, and apparently my memory sockets were in good condition, 'cuz the system works fine now with no gui issues.

may be testing the currently extracted ram on more aggressive linux software later on, on a diff. laptop.

 

[Cerb]

A positive result means something is bad, but a negative result does not ensure that nothing is wrong. Passing a memory test when RAM appears to be the problem indicates it's likely not RAM, but no RAM test is perfect.

One may find errors another doesn't, based on how it tests. I've seen it every which way, between Memtest86+, Windows' RAM diag, and Dell's RAM diag.

 

[ArisVer]

Off topic question.

How does an ECC buffered stick and a normal stick compare in terms of a bad memory module? Would an ECC buffered stick continue to work partially?

 

[Cerb]

Off topic question.

How does an ECC buffered stick and a normal stick compare in terms of a bad memory module? Would an ECC buffered stick continue to work partially?

If the error could be caught and corrected, yes (typically, all 1-bit errors can, and most 2-bit). You still need to replace the module, however, because they tend to get worse over time, meaning uncorrectable errors before the machine is decommissioned are possible. It's nice that ECC can correct some errors, but the checking is the really valuable part.