[xutnubu]

Hello, fellows, how does it look?

Max load voltage is 1.128v @ 4.3GHz, overclocked using the Turbo multipliers.

Here's the Prime95 result. I used the Blend test recommend in the ASRock OC thread.

8 Hours:

Spoiler: Warning: Spoiler!

Taken 5 mins before ending:



Taken right after it ended:

HWiNFO64 readings:

Spoiler: Warning: Spoiler!

I've also play some hours of BF3, it seems stable. On that game it stays at around 60c.

There are some things that make me worry a bit.

The 2nd core seems to always be 4c hotter than the other cores. Is this why it did fewer (170 less) tests than the other cores?

Also, are the number of tests ok for 4.3GHz?

Also, I've seen in different places that people need more voltage than I did here for 4.3GHz. How bad it is to run your CPU at low voltages, can something go wrong?

Any tips are welcomed.

 

[justanoldman]

There will always be difference between your cores, 4 to 5c is not unusual at all. A 10c difference however would be cause for concern.

Lower voltage is a good thing, a very good thing actually. You could have a good chip there (good is measured by how little voltage is needed for each multiplier), but you would have to try for 4.5 or above to find out.

Unfortunately your cooler will limit you going much further, you will need to upgrade it if you want to try for a higher overclock - or delid your chip, but only if you can afford to make a mistake.

 

[justanoldman]

Also, there are a huge variety of opinions on this but I personally do not feel 8 hours is enough to find stability. It is sufficient if you just game, email, surf, etc. and can afford the occasional crash or instability.

If you want a higher level of stability then run 24 hours of Prime95 custom torture test with the memory used changed to 90%+ of available with no errors, no signs of instability, and no occurrences in your Event Viewer->Custom Views->Administrative Events. After than you can run IBT on max, game, and try any program you can think of to see if you have issues.

 

[xutnubu]

Quote:

Originally Posted by justanoldman

There will always be difference between your cores, 4 to 5c is not unusual at all. A 10c difference however would be cause for concern.

Lower voltage is a good thing, a very good thing actually. You could have a good chip there (good is measured by how little voltage is needed for each multiplier), but you would have to try for 4.5 or above to find out.

Unfortunately your cooler will limit you going much further, you will need to upgrade it if you want to try for a higher overclock - or delid your chip, but only if you can afford to make a mistake.

I definitely can't.
Quote:

Originally Posted by justanoldman

Also, there are a huge variety of opinions on this but I personally do not feel 8 hours is enough to find stability. It is sufficient if you just game, email, surf, etc. and can afford the occasional crash or instability.

If you want a higher level of stability then run 24 hours of Prime95 custom torture test with the memory used changed to 90%+ of available with no errors, no signs of instability, and no occurrences in your Event Viewer->Custom Views->Administrative Events. After than you can run IBT on max, game, and try any program you can think of to see if you have issues.

I only use it for gaming.

Maybe I can do that test later. My brother is desperate to play, so I can't right now

 

[smex]

Quote:

Originally Posted by justanoldman

... A 10c difference however would be cause for concern.
.

No, even 10 can be OK! you can try to reseat it and get better temps, but there will always be a difference. as long as it isnt much over 10°C its ok.
the heatsink does not always make equal preassure on the cpu and the TIM has not the same thickness everywhere. this is one of the reasons for the temp
difference. further.. the core that is beeing used is producing more heat than a currently unused core.. thats the magic behind all of this.

 

[justanoldman]

Quote:

Originally Posted by smex

No, even 10 can be OK! you can try to reseat it and get better temps, but there will always be a difference. as long as it isnt much over 10°C its ok.
the heatsink does not always make equal preassure on the cpu and the TIM has not the same thickness everywhere. this is one of the reasons for the temp
difference. further.. the core that is beeing used is producing more heat than a currently unused core.. thats the magic behind all of this.

I will respectfully disagree. If you run Prime95 for a few hours, and have a 10c difference between max core temps on an Ivy chip then chances are you do not have your TIM and cooler optimally installed. Finding the optimal amount and placement of your TIM, and optimally installing your cooler with regard to orientation, and pressure takes a few hours of testing. It is worth it in my opinion though, and I think too few people take the time.

 

[smex]

Quote:

Originally Posted by justanoldman

I will respectfully disagree. If you run Prime95 for a few hours, and have a 10c difference between max core temps on an Ivy chip then chances are you do not have your TIM and cooler optimally installed. Finding the optimal amount and placement of your TIM, and optimally installing your cooler with regard to orientation, and pressure takes a few hours of testing. It is worth it in my opinion though, and I think too few people take the time.

you did not disagree

you said the same as me except the possibility of different temps due different loads. what you are saying if alle 4 cores are utilised in prime, over time this big difference
is too big, and in this scenario you are right.. @ 100% load on all cores.
but when gaming or using apps, an app might use just one core or two, and theese two will have a higher temp then the ones that wasnt used.

see, so you absolutely dont have to disagree

 

[justanoldman]

Quote:

Originally Posted by smex

you did not disagree

you said the same as me except the possibility of different temps due different loads. what you are saying if alle 4 cores are utilised in prime, over time this big difference
is too big, and in this scenario you are right.. @ 100% load on all cores.
but when gaming or using apps, an app might use just one core or two, and theese two will have a higher temp then the ones that wasnt used.

see, so you absolutely dont have to disagree

Agreed, you will definitely have higher differences when doing normal activities. The OP showed us his Prime95 temps so that is what I was referring to, so you are right and I agree with your agreement that we don't disagree about our disagreement.

 

[xutnubu]

Bump.