The new Intel Core i9 10900K can get pretty hot under heavy load and consume quite a bit of power if pushed too hard Caseking tip viner, overclocker, and general CPU prodder Der8auer has posted a video showing the Comet Lake processor, posted a video showing the performance differences between the top 30 versions of the processor.

Even within the 10-core range of 10th generation CPUs, the silicon lottery is indeed alive and uncompromised. You may be super lucky and get an i9 10900K chip that is cooler and less thirsty than others right out of the box. If that happens, you'll be able to get decent results with overclocking.

Conversely, you might end up with a dud that is already running at 92°C at stock settings or a super leaky processor that consumes 20W more than average. Either way, the overclocking fun will soon be over. Or there will be no hope of building a powerful small form factor gaming rig.

And you won't know which you've chosen until you take it home, plug it in, and fiddle with the various performance knobs. Or unless you splurge and buy a pre-binned version of the chip, like Caseking's.

Binning is the process of testing a batch of CPUs (or other high-performance silicon) to see how different versions of what should be the same chip actually perform. Vendors sell pre-selected versions of a CPU that they know will outperform others, for example, at a hefty premium.

Der8auer tested 30 different 10900K processors from the exact same production lot, in the exact same system, with the exact same settings, and at the same room temperature. He found that with all chips running at 5.1 GHz and 1.25 V, six had peak temperatures below 80°C, 12 had average temperatures above 82°C, and one extreme anomaly recorded 92°C

.

He also tested the peak power consumption of the packages, recording 232 W for the lowest and 278 W for the highest.

In both cases, there is a difference of 13°C in peak temperature and 46 W in peak power; on average across 30 different chips, the temperatures and power consumption are not that bad, but the little K at the end of the CPU code does not mean much when the silicon lottery yields bad results. Nominally it may be overclockable, but if it reaches 92°C on normal settings, you'll struggle to get much more performance out of it.

And remember, this is the same batch of chips, probably manufactured at the same time; as De8auer points out, Core i9 CPUs from different lots may have even more variability.

Another point worth noting in De8auer's study is that the coolest 10900Ks were not the lowest power consuming, nor were the highest leakage or highest power consuming chips the hottest.

Looks like we got lucky. Our Core i9 10900K may be a little higher than average temperature-wise, but it seems to be a little less thirsty. This may be why we were able to comfortably achieve 5.3 GHz on all cores without melting.