Down the rabbit hole: (a little off topic)

[Pokey]

“ram may be vulnerable to high frequency row hammer bit flips”

I bet Rudy could turn that into a “believe it or don’t”.

I had a rig that was spontaneously rebooting so I tried some things but it boiled down to either ram or psu. I downloaded the latest Memtest86 and proceeded to test the ram modules. I got the above message and freaked out a little as I had no idea what that meant. So I read up on row hammer bit flipping. Then realized the message did not mean it was actually happening since there were no errors being reported. The key here is the “may be”, So the ram appears to be behaving. The psu tested good but I replaced it anyway and all is right with the world now. I guess at some point after being on a while the voltage must drop enough to crap the rig out.

Only took me 2-3 days to figure it out…………………… :\

 

[TennesseeTony]

“ram may be vulnerable to high frequency row hammer bit flips”

I can verify it's vulnerable to hammers, even at low frequency, causing back flips.

Glad you got it sorted out.

 

[Drsignguy]

What Brand PS/wattage and what brand/wattage did you replace it with? It is always interesting to hear all the different types of problems that are experienced with the wonderful world of building a PC. Especially a DC PC.

I too am glad to hear you solved your problem!

 

[Rudy Toody]

I can verify it's vulnerable to hammers, even at low frequency, causing back flips.

The damage depends on the size of the hammer and whether you are still holding it while you do the back flip.

 

[Pokey]

What Brand PS/wattage and what brand/wattage did you replace it with?

I got a Corsair CX750 80 Plus Bronze. It has a single 12V rail. I have been leaning toward the single 12V rails of late. Plus it has four pci-e 6/8 cables. Good for two GPU cards.

 

[Pokey]

The damage depends on the size of the hammer and whether you are still holding it while you do the back flip.

LOL, I knew you could do it................. thanks.

 

[VirtualLarry]

The damage depends on the size of the hammer and whether you are still holding it while you do the back flip.

Is that a new Olympic sport? "ROW HAMMERING"

 

[Pokey]

The problem had me thinking about getting hammered.

 

[Drsignguy]

I got a Corsair CX750 80 Plus Bronze. It has a single 12V rail. I have been leaning toward the single 12V rails of late. Plus it has four pci-e 6/8 cables. Good for two GPU cards.

The corsair Power supplies are extremely reliable and pretty much bullet proof. I would really like to know what a "bit flip" is :\.....As for a voltage drop that most likely caused your issues, your Vid cards were the draw as the PS was on its last leg...And I bet you caught it before it could have done some major damage.....nice work babysitting pokey

 

[TennesseeTony]

Speaking of off topic...

Weekly stats will be delayed for about 8 more hours...stuck at work.

 

[Rudy Toody]

I have protected my computer from the hammering bit-flips by wrapping the memory sticks in tinfoil. Now all I have to do is find a new hat.

 

[Pokey]

I have protected my computer from the hammering bit-flips by wrapping the memory sticks in tinfoil. Now all I have to do is find a new hat.

I'm sure that will work, but I can't help you with the hat.

 

[Assimilator1]

Being a mechanic, I have a wide choice of hammers you can borrow!

 

[Pokey]

My PSU fix did not hold. I got a couple more spontaneous reboots so I replaced the, relatively new, Crucial memory modules with new GSkill modules and it has been four days now of steadiness. I am right at declaring the problem fixed.

I have had good luck with Crucial in the past, but now I might just become a GSkill user.

The good news is the old PSU may be able to come off the bench sometime in the future.……….

 

[Assimilator1]

Hope it's sorted , btw you could of narrowed down which RAM module is causing you problems by removing all but 1 module & then testing them in sequence.

Somewhat related, I bought an extra 4GB of RAM for my 2nd rig, OCCT tested the rig & it failed in 15 mins, thinking it could be the o/c settings being too much for 8GB of RAM I did various tweaking, but to no avail. Removed the 2 new sticks & returned settings to original, still failed! Turns out what was (~2yrs ago) a 24hr stable OCCT o/c is no longer! Fails after ~1hr , bumping vcore increased that to nearly 5hrs, just testing with a 1MHz drop in FSB atm, once passed, then onto testing with 8GB of RAM .