noob to overclocking but surely somethings wrong i5 3570k



seems to be running 5'c cooler now

thats with the turbos set to 4.2ghz
pll @ 1.70v
watts/amps limit still at 250
vcore @ normal
with the divd -0.070
and LLC on low

when i change the voltages it doesnt seem to affect the showing fiqure on coretemp

but does affect the cpu-z fiqure when i change the divd
think im going to drop it to -0.080
and then leave the stress test on prime running for awhile

massive thanks to wazza for being patent enough with me haha
 
coretemp might be off,usually cpu-z is right

you can try for 4.4ghz if you want,depends on temps ect,the lower the voltage you can get stable the better cos it will be less hot
 
My thermal paste is about a year and half old now and it was only the original crap that comes with the matterhorn
So ive just bought some as5 paste and a cleaning kit so when I come ill see if that makes a difference
As its ment to be a good fan
 
Hi Chris, just to give you my settings as I have the same CPU/motherboard set up.
My CPU is a fairly poor clocker as I need 1.32v to get 4.4ghz stable.4.5ghz is 1.356v.
I've set mine now and left it at 4.3ghz as the voltage needed for a stable system works out as 1.27v
As you can see above 4.3ghz I need a lot more voltage and that means more heat !!!!
At 4.3ghz I prime stress, my CPU reads about 70*c, gaming, nothing higher than 60*c, this is with an H80 cooler , your cooler should be adequate for this sort of o/clock.
PWm set to high performance
Core volt response set to fast
Load line medium
Dynamic vcore -0.040
CPU clock ratio 43

Everything else set on auto, been like this for well over 12months

This is just a guide as every CPU is different
Once you reapply your thermal paste you should see a few degrees drop

Or you could delid!!!! Haha
 
Christurbo said:
when i change the voltages it doesnt seem to affect the showing fiqure on coretemp
Click to expand...

The thing you see in CoreTemp is the VID of your processor, or the default voltage Intel has tested and set this particular sample to work on. I generally would use Real Temp for temperature monitoring, it is just a better tool on my opinion. HWInfo64 is also amazingly functional free portable program that I use every time I need monitoring information.
The voltage you see in CPU-Z is the current voltage measured by the sensor inside the chip. It is more or less accurate and can be generally trusted.
Christurbo said:
My thermal paste is about a year and half old now and it was only the original crap that comes with the matterhorn
So ive just bought some as5 paste and a cleaning kit so when I come ill see if that makes a difference
As its ment to be a good fan
Click to expand...

Why do you think the original "crap" that came with the matterhorn is worse than Arctic Silver 5 which is ancient, like 10 years old when formula goes. There was a very nice thread about thermal compounds with links to reviews, methods of application and everything. Even if you have gone with the popular here MX-4 it still would have been better the AS5.
Christurbo said:
but does affect the cpu-z fiqure when i change the divd
think im going to drop it to -0.080
and then leave the stress test on prime running for awhile
Click to expand...
If the dvid is at -0.07V and the voltage under load is 1.25V, I think you can easily try -0.17V. :D I mean 1.25V is extremely high for 4.2GHz, CPUs usually do that at say 1.1-1.15V. There are samples that do 4.5GHz at 1.15V. Go for bigger steps incrementally, say try - 0.12V, if it does no fail and passes your stress tests, lower with same quantity. When it fails, start putting it up in small increments, say +0.01 or +0.005. I hope I have explained this understandably - go for -0.05V drops until it fails, then from the last dropped point start increasing by +0.01V/+0.005V until it is stable. If every single step lower than your previous setting fails this will bring you back to your original stable value.
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom