Upped the Vcore 2 more notches and temps reached 90C
stopped it before it got any higher... is 90 OK? 10 more degrees and it would shut down
Hi nicnac1,
90°C is hot . . . . very hot!
For testing purposes that's fine but for normal day to day use you don't want your chip burning away like that for long. We were just trying to see if the chip became stable with the increase voltage but the temps are doing us no favours!
Most folk that are really into their overclocking normally go to great length to keep their chips/&PC as cool as possible because the hardware seems to perform better that way. I don't know the scientific/engineering reasons for this exactly but it's roughly to do with the many electrical pathways and circuits in the chip don't perform as well once they get near boiling point!
No two chips are alike, the silicon reacts to voltage and heat differently. Intel rigorously test the silicon in their labs to see if it performs within certain parameters and based on that testing mark the chip up as a certain product and program it with a suitable [VID].
Some silicon cannot run high frequencies once a certain temp threshold is broken, it may be able to run 3.6GHz while the load temp is kept below 65°C but once this figure is breached the circuits don't perform very well and you get errors.
The rub here of course in your situation is your kinda caught in a pincer movement, it's something I call the
Hot Gates, i.e your chip appears to be needing two things that work against each other
- It can't make the clock if it doesn't have enough vCore
- It can't make the clock if it runs to hot
While adding more vCore is needed for chip stability the added heat maybe causing instability, it's a pain!
Improving your cooling is one way of trouble-shooting this problem, swapping out chips is another (or sending it for testing). Another handy method to work out what is going on is to perform some
CPU Voltage Scaling Tests . .
What that means is, starting at default/stock speeds you work out exactly how much vCore a chip needs to run stably at a given frequency. As the Frequency rises you normally find you need to add more juice and over the period of testing a pattern emerges where you find out the sweetspot voltage wise of your chip. You find the lower freqencies respond well to a slight bump in voltage but as the MHz ramp up it starts needing greater and greater doses of voltage to even climb a measly 20MHz. At this point you know you are well-past your chips sweetspot!
Although
CPU Voltage Scaling testing takes time it is a very useful process, you get to know *exactly* what vCore a chip needs at a given frequency and witness how this voltage effects temps. If you have a Energy-Brick/Power Meter attached you would also get to see how the added vCore effects the power draw from the wall (i.e your PC's running costs ££).
I think it would be a handy process for you to perform and buys us some time while you organise your improved cooling.
What you need to do is reset your BIOS back to default 2.4GHz (9x266) and leave all the Voltages on [Auto] except vCore, you then slowly but surely keep turning vCore down, down, down with a bit of quick testing inbetween until you find the correct voltage. I want to know what is your chips True [VID]. You will find that if your chip doesn't have enough juice it will either fail Prime or your PC will reboot, if that happens just nudge up the vCore one bump and test again.
Once you found out what it needs to run stable at 2.4GHz we can progress quickly to 3.0GHz!
