AMD 1700 Precision Boost Questioning.

Intel doesn't boost all cores either. Both Intel and AMD initially boost only a single (Or two, depending on the generation) core to a specified max boost clock. More cores will boost if needed but the more cores that boost, the lower the maximum boost clock that is applied for thermal reasons. Generally, automatic overclocks AND standard boost designs use higher voltages than you or I would do if we were manually overclocking because we could incrementally adjust multipliers and voltages until we settled on a satisfactory maximum based on temperature/stability.

For the sake of the silicon lottery

(Every processor ever made will have it's own specific characteristics and none will behave exactly the same at identical clocks or voltages, hence "lottery" as some chips will simply be able to clock higher at lower voltages than others. There is no way for AMD or Intel to account for this aside from simply lowering the margins for all with a few caveats based on thermal reactions)

, and since there is no way for them to automatically do as we would do manually, MORE voltage than what is really necessary for a core to be stable at a given clock speed is applied by OC utilities (OEM and otherwise) and turbo/precision boost (or turbo core) designs so that they can err on the side of stability rather than what is most efficient thermally. The problem with that is, once more than one core is boosting that maximum core clock per core has to be lowered in order to reduce the overall package temperature, unlike a manual overclock where we will have already determined at what clock speed and voltage ALL cores can be simultaneously increased to without exceeding Tjmax/Tjuncture temperatures at full load.

SO in reality, that does not really look all that unusual and there is a good deal of information available on this very subject.

You start at the maximum boost clock, because thermals are still good and power draw isn’t very high with only two active cores. As the benchmark carries on, those two cores boost to 4.1GHz using XFR, but only briefly. As the chip heats up and more cores get loaded up, the clock speed drops to 3.7GHz, before lowering again to the chip’s base clock of 3.6GHz.

Click to expand...

A very detailed explanation of the Ryzen boost implementation is available here, but I assure you that while there are SOME differences in the way AMD and Intel go about this, the gist of the practice is the same at it's essence:

http://www.nag.co.za/2017/03/06/a-brief-comparison-of-amds-precision-boost-versus-intels-turbo-boost/