harly wrote:
So who has information on the consequences so that risk can be weighed?
Lets say I keep running this as my 24/7 and I become one of those people that bust their CPU. Statistically, what is the likelihood that the damage will be confined to the CPU? If it is highly likely that only the CPU will stop working and just need to be swapped out, then I'm reasonably happy with that risk. But if there's say 50% of cpus that fail due to my level of overclocking also break the mobo, then it becomes somewhat more expensive!
1) Intel performs the risk assessment when they define the maximum VID and VCC thresholds at stock frequency. Anything past that is in the hands of the user.
2) Nobody has the info you want. Most people tend to use a certain level of common sense. Some don't, and push too far, then report their case back on forums.
3) If you are going to a push a system, it's always wise to cool all current delivery and consuming components as best you can. That goes from VRMs, to the chipsets, memory and also the CPU. Understand the relationship between voltage, frequency (that goes for all sections of the processor) and current. Don't go too far over stock VID (actually VCC if you take LLC into account) or frequency and you limit your chances of failure somewhat.
4) As with all things you may end up with a weak sample. Only Intel knows how many failures are experienced on non-overclocked workstation systems that are built exactly to Intel tolerances. As overclockers push further than that, they are playing with any failure rate buffer margin Intel allow themselves when they determine optimal power envelopes.
5) When a high current component fails or any component that passes current for that matter, there is always a risk it will take surrounding components with it.
-Raja
.jpg "Raja"){kind=avatar}
