32GB of B-die, 3200 C14, overlocking guidance?

Associate
OP
Joined
8 Nov 2006
Posts
409
Location
Scotland
P2C7By3.png

Still working away at this lol, memory stability testing really is the dark arts. Had to start over a few times and instead have been slowly doing settings one by one, going for a 3 cycle test for a quick rule out of really unstable and then doing a 9 cycle overnight.

Worked my way back to trying to secure CL14 at 3800. I gave in on CL14 quite quickly, but with some advice from around the web have been trying to get here again.

I've got my RAM maxed out at the highest voltage I feel safe with, 1.55v, with active cooling (fan cooled), but hoping to drop that voltage a bit if I can. I'd rather be at 1.5v, but lets see if I can get CL14 fully stable first.

RAM temps during stability tests around 35~39 degrees which I'm OK with. High 40s, or, the nearer you get to 50 degrees, seems to be where b-Die starts to error out or have issues.
 
Soldato
Joined
6 Jun 2008
Posts
11,618
Location
Finland
I've got my RAM maxed out at the highest voltage I feel safe with, 1.55v, with active cooling (fan cooled), but hoping to drop that voltage a bit if I can. I'd rather be at 1.5v, but lets see if I can get CL14 fully stable first.

RAM temps during stability tests around 35~39 degrees which I'm OK with. High 40s, or, the nearer you get to 50 degrees, seems to be where b-Die starts to error out or have issues.
Remember to do final stability testing with load also on CPU and GPU.
GPU load rises ambient temperature inside case and that can push memory too warm.
And CPU load is needed to make sure IF is stable and that heat from cores won't affect memory controller.

Because of that extra heat settings which might be stable in long memory tests could crash during gaming.
 
Associate
OP
Joined
8 Nov 2006
Posts
409
Location
Scotland
Remember to do final stability testing with load also on CPU and GPU.
GPU load rises ambient temperature inside case and that can push memory too warm.
And CPU load is needed to make sure IF is stable and that heat from cores won't affect memory controller.

Because of that extra heat settings which might be stable in long memory tests could crash during gaming.

Yes this is a good point. I'm running active cooling on the RAM right now so its fine but depending on what voltage I land on I might see if it can make do with air flow in the case.

RAM testing truly is the dark arts, the pain you feel when you get 1 error in TestMem5 overnight during an 8 hour cycle.

So, I'm still working on where I will land first. CL16 is stable and fairly tight at 1.4~1.45v depending on some secondaries. CL14 is proving to be a rollercoaster trying to get stable.

I gave up on GDM off and CL15 after a day or so, can boot but it's truly the devil trying to get stable at 1T.
 
Back
Top Bottom