SOLVED Modestly overclocking my Phenom II X4 965 Processor

[pqwoerituytruei]

I have my Phenom II X4 965 Processor on a ASUS M4A79XTD EVO with a Hyper 212 Evo on top of it, my temps top out at 44C at the stock 3.4GHz (prime 95)
Every part with pictures
I have never overclocked before I was going to try the auto overclock, but i don't want to overclock my ram which it wanted to, it is already DDR3 1600 which is plenty
I tried to set the CPU to 3.6GHz (1st thing above 3.4GHz) and leave the rest on auto, that did not work (Edit: what i set was actually the north bridge, not the CPU freq)
I am doing this from my BIOS
Main reason for overclocking it is to see how it is done, so i was not going to overclock it much, definitely not going for 3.9 or 4 GHz (yet), i was thinking 3.5 or 3.6 GHz, i assume i have to use 100Mhz intervals

 

[keny]

The easiest way for you to grab a quick fast overclock is to just bump the CPU multiplier (stock x17) up a notch and try and run prime, if it fails bump up the CPU vcore (CPU voltage) a notch until it will pass and just repeat those steps until you are happy with where you want to be or the heat of the CPU gets over 55c under load (prime).... Easy

 

[trents]

Run the Prime95 blend test for 20 minutes in the initial stages of your overclock but when you seem to have maxed it out run the stress test for at least two hours to confirm stability. If you pass it for that length of time it's pretty reliable. Failing it means blue screen, spontaneous restart, lock up or one or more of the four Prime95 core workers drops out. Core temps not exceed 55-60c and CPU socket temps should not exceed 70c. Whenever you stress test have HWMonitor open on the desktop to monitor those temps.

 

[pqwoerituytruei]

just set it to 17.50 and now have 3.5Ghz, i have a hrad time believing that will not be stable with my setup, wonder if i can get over 44C now with my temps
i think 3.6 should be just as easy since they sold phenom II chips at 3.6GHz
Would a small overclock like this be just as easy on a Athlon II?
this is what my sensors are reporting for the voltage

Adapter: ACPI interface
Vcore Voltage:        +0.92 V  (min =  +0.85 V, max =  +1.70 V)
 +3.3 Voltage:        +3.45 V  (min =  +2.97 V, max =  +3.63 V)
 +5 Voltage:          +5.05 V  (min =  +4.50 V, max =  +5.50 V)
 +12 Voltage:        +12.60 V  (min = +10.20 V, max = +13.80 V)

if the voltage core is set on auto will it bump it up as needed or will it fail, i would rather not touch the voltage since that will increase the wear on the chip

 

[Felixeab]

Hmm, I had one myself, and they are very prone to instabillity with low votage. Be sure it gets enough power, and try not to alter the FBS (Memory Clock, set at around 200ish?) or the Hyper-Transport crashes the system. It must stay at 1000Mhz. Do not go below that, though you can go a little above it with some voltage bumps, but that means you need good cooling as well.

 

[trents]

Hmm, I had one myself, and they are very prone to instabillity with low votage. Be sure it gets enough power, and try not to alter the FBS (Memory Clock, set at around 200ish?) or the Hyper-Transport crashes the system. It must stay at 1000Mhz. Do not go below that, though you can go a little above it with some voltage bumps, but that means you need good cooling as well.

Felix, I think you are confusing the Athlon II with the older Athlon 64. The Athlon 64 had a native HT Link of 1000 mhz but the Athlon II has a native HT Link of 2000 mhz. And the HT Link is not the same as the HT Reference. Hyper-Transport is a rather vague term unless you qualify which dimension of it you refer to.

pqwoerituytruei, overclocking the Athlon II is a very different animal because the core speed multiplier is upwardly locked. You would need to use the FSB (technically, the HT Reference) to overclock and that gets complicated because some other frequencies are tuned to it and must be manipulated independently to maintain stability of the system.

 

[pqwoerituytruei]

thanks, sounds like more trouble than its worth

anyway my new max temp seems to be 42C up 100Mhz and down 2C
that is eigther the thermal paste setting in or the the season changing to winter

 

[keny]

You're temperatures on you're current overclock should not go any higher than they are at stock until you raise the vcore, faster clock speeds don't make it hotter, its the extra voltage applied to get the faster clock speed that causes the heat, and yeah they did make higher speed phenoms but they were higher speed due to them being of a higher quality when produced and could make that higher clock speed at the same voltage as the lower clocked phenoms, it doesn't mean all phenoms can achieve the higher binned phenoms speeds, otherwise why would you buy the higher speed chips!! there's only one way to find out how good you're phenom is

 

[trents]

keny, my experience is that faster clock speeds do generate some rise in temps but by far the biggest heat producer is increasing voltage. Some thermal pastes do "cure" with time, like Arctic Silver 5.

 

[keny]

I will take you're word for that Harry but it must be by a tiny amount as I've never noticed it myself, and to the OP please don't be afraid of raising the vcore, you're CPU will last a long time under 1.5v and 55c core temp

 

[pqwoerituytruei]

i don't plan on upgrading my motherboard for a long time and since i don't have a am3+ socket for Vishera (or the next cpu reversion), i don't want to not be able to replace the cpu since it will be discontinued by then
using the stock coolermaser paste
would auto on the vcore adjust to accommodate the multiplier? or does it auto to the chips stock setting
roughly what are the odds of 3.6Ghz at stock voltage working on this chip
oh my vcores is different when under full load according to my sensors

Vcore Voltage:        +1.29 V  (min =  +0.85 V, max =  +1.70 V)
 +3.3 Voltage:        +3.44 V  (min =  +2.97 V, max =  +3.63 V)
 +5 Voltage:          +5.05 V  (min =  +4.50 V, max =  +5.50 V)
 +12 Voltage:        +12.58 V  (min = +10.20 V, max = +13.80 V)

during idle my cpu sites at 800Mhz which is which explains the difference

 

[trents]

Many people get 4.0 ghz or a little better on that CPU series. I would expect you should be able to attain that with your supporting system components.

 

[pqwoerituytruei]

true, but how far with stock a stock V core voltage, at this time i don't want to mess with the voltage, if i was overclocking for a good reason that would be different

 

[keny]

Every cpu is different and the only way to find out is to try it for yourself, you may find it fails after one bump on the multi or it may go to 20x on the multi at stock vcore, its a suck it and see problem that only you will be able to find out.

 

[trents]

true, but how far with stock a stock V core voltage, at this time i don't want to mess with the voltage, if i was overclocking for a good reason that would be different

As long as temps stay within a safe range, increasing voltages will not damage anything unless you go to extremes like in excess of 1.6 vcore where electron migration might start occurring. When people come to OC Forums we kind of assume they are wanting to push the envelope, hence the name of the forum.

 

[pqwoerituytruei]

this means it is not stable right?

[Wed Nov 14 11:27:42 2012]
FATAL ERROR: Rounding was 0.5, expected less than 0.4
Hardware failure detected, consult stress.txt file.

 

[keny]

Yep! That's not stable, needs more vcore

 

[pqwoerituytruei]

what is the stock vcore, it is on auto right now, i backed it to 3.5GHz and running prime for a while to make sire that is stable
almost 30min so far no error 44C (ran about a hr 10 min no error then i stopped it)

 

[Special7]

You should be able to do 3.6GHz on stock volts or even under-volted a bit like me, I would go further but I am limited with this stock cooling setup.

 

[pqwoerituytruei]

are you sure it is stable?
one thread on while running prime crashed for me after 10 min or so
what is the stock voltage, my board might me under volting it as it for all i know
highest vcore i have seen from lm-sensors is 1.310 or was that 1.312, just under 1.3 while using prime