• Some users have recently had their accounts hijacked. It seems that the now defunct EVGA forums might have compromised your password there and seems many are using the same PW here. We would suggest you UPDATE YOUR PASSWORD and TURN ON 2FA for your account here to further secure it. None of the compromised accounts had 2FA turned on.
    Once you have enabled 2FA, your account will be updated soon to show a badge, letting other members know that you use 2FA to protect your account. This should be beneficial for everyone that uses FSFT.

I7-3820 decay

M76

M76

[H]F Junkie
Joined
Jun 12, 2012
Messages
15,275
I currently have an I7-3820 cpu that runs at 4375Mhz I had to lower the frequency from 4500, because I started to have stability issues. When it was new I could run it at 4625. The CPU is now two years old. Should I be worried? I'm not used to this kind of decay all the cpus I owned before didn't loose their overclockability over time at all.

Maybe I ran it too high voltage? What is acceptable for these cpus?


If I decide to sell it what would be a good upgrade? The 5960X and the 5930K are not in my budget. So the only options are 4930K, or 5820K, or wait for 6700K.
 
What're you using to cool the thing?
It's possible that your CPU cooler needs a thermal paste refresh.

And/or if you're using a liquid cooling system, you need to change out the coolant.
 
I changed the cooler about a year ago. That lowered the core temperatures by 10-15 degrees, but it didn't allow me to return to a higher oc. It's a closed loop AIO. CM Nepton 140XL.
 
what voltages was you using to achieve the past OC?. what you describe is tend to be a degradation, that can point to high temperatures over long periods and/or high voltages over long periods.

What kind of stability issues are you getting actually? there are another things to check as RAM and BLCK.. Overclock, IIRC in those chips can be achieved by the partially unlocked Multiplier or BCLK, the BCLK put extra strain in the motherboard so is a thing to check..

and well, you can try just default motherboard settings, clear CMOS and start the OC process again to discard any bug related with the mobo..
 
Araxie said:
what voltages was you using to achieve the past OC?. what you describe is tend to be a degradation, that can point to high temperatures over long periods and/or high voltages over long periods.

What kind of stability issues are you getting actually? there are another things to check as RAM and BLCK.. Overclock, IIRC in those chips can be achieved by the partially unlocked Multiplier or BCLK, the BCLK put extra strain in the motherboard so is a thing to check..

and well, you can try just default motherboard settings, clear CMOS and start the OC process again to discard any bug related with the mobo..
Click to expand...

BCLK is the same at 4375 as with 4.5 or 4.625. Ram settings are unchanged as well, only the multipler is different and I lowered the voltage by some degree. But I don't have the exact numbers now.
 
check the numbers and tell us, the stock BCLK for that chip is 100mhz*38 CPU multiplier. for a total of 3800mhz. you are probably overclocking with BCLK so, try reverting BCLK to 100mhz and just increase the CPU multiplier, IIRC the max for that chip is 43 try at that speed stability and voltages..

A friend of mine have that chip at 4625mhz with 125mhz*37 CPU multiplier and 1.36V. using 38 multiplier for 4750mhz ended requiring above 1.4v which ins't good for long term usage.
 
Araxie said:
check the numbers and tell us, the stock BCLK for that chip is 100mhz*38 CPU multiplier. for a total of 3800mhz. you are probably overclocking with BCLK so, try reverting BCLK to 100mhz and just increase the CPU multiplier, IIRC the max for that chip is 43 try at that speed stability and voltages..

A friend of mine have that chip at 4625mhz with 125mhz*37 CPU multiplier and 1.36V. using 38 multiplier for 4750mhz ended requiring above 1.4v which ins't good for long term usage.
Click to expand...

As I've said the BCLK is 125 at 4375 as well, so it has no effect on stability. The only factor is the CPU. The multipler doesn't go over 38 so I can't OC it without BCLK at all. If I set multipler to anything over 38 it will just boot at 3800.
 
Things that decay and need replacing:

Thermal paste.
Fans
Power Supply as voltages will start to waver over time.

Things less likely to decay:
Motherboards. VRM temps increase because of old thermal pads, capacitors get old.
Video card memory dies because of overclocking

Never seen a CPU degrade over time; only the components around it.

Lastly, have you tried reseating the AIO cooler? This may sound like I'm saying user error, but it is really easy to install one without torquing it properly and leaving one side slightly misaligned. Here is a video on how to do it. Skip to 2:30. Main part idea to get from it is to torque it down using a diagonal method.
 
I set the clock to 4.5Ghz again and now it seems stable. It turns out two of my case fans weren't running causing excess heat in the case. And since the CPU cooler is set to a fixed RPM it didn't counteract the extra heat.

But the VCORE is 1.45 which is a bit too much, but it's not stable with less. But it was working like this for 2 years, so I think I'm ok. Even if it fails I'll just have one more reason to upgrade since right now I can't convince myself.

Temps are around 45 without load, and goes up to 80 running prime95
 
As I mentioned I've found the culprit. The fans wouldn't spin up, so I had to give them a little more power from bios.

It doesn't matter if it's clockwise or counter clockwise. You shouldn't move around either way. What matters is you always move to the screw across from it not next to it. And you shouldn't tighten the screws fully at once just go bit by bit after it starts to tighten, maybe only quarter turns.
 
cageymaru said:
Things that decay and need replacing:

Thermal paste.
Fans
Power Supply as voltages will start to waver over time.

Things less likely to decay:
Motherboards. VRM temps increase because of old thermal pads, capacitors get old.
Video card memory dies because of overclocking

Never seen a CPU degrade over time; only the components around it.

Lastly, have you tried reseating the AIO cooler? This may sound like I'm saying user error, but it is really easy to install one without torquing it properly and leaving one side slightly misaligned. Here is a video on how to do it. Skip to 2:30. Main part idea to get from it is to torque it down using a diagonal method.
Click to expand...

I hadn't seen one decay either, until my 920-D0. Hummed along at 4.2 for ~3.5 years with good cooling, normal voltages, etc. Started getting random crashes (even when trying to boot; codes related to cpu), re-seated my cooler multiple times, cleaned pins and pads, etc. Got a bit better at 4.0 for a while until it became more frequent again. When it started crashing more often at 3.8, I ordered a 6-core Xeon. Been rock solid at 4.2 for a year now.
 
I've never heard of CPUs degrading unless they were run above their absolute maximum voltage or temperature for a long time. OTOH I have heard of degradation of thermal paste (dries out), fans (slow from dirt or oil breakdown causing more friction), heatsinks (coated with dust), and capacitors (conventional wet ones quit filtering as well). Also is the room temperature hotter now?
 
You must log in or register to reply here.
Back
Top