Did you look at what is happening with the 8700's clock under load? There is some evidence that less expensive B360 boards may not have sufficient power capacity to sustain the all-core max turbo of the upper end Coffee Lakes. From TechSpot: "Everything looks great for the first 30 seconds, the 8700 holds 4.3 GHz on all 6-cores and we ripping through this workload. Then like that the motherboard winds the 8700 back down to 3.2 GHz for 7 seconds before cranking back up to 4.3 GHz for 8 seconds and then repeating the process. The reason this is happening is because the motherboards VRM is overheating, or at least reaching the thermal limit, it then backs off the Core i7-8700’s power draw, cools a few degrees over a 7 second period and then ramps power delivery back up until the thermal limit of the VRM is once again reached, which as I just said took about 8 seconds." Not necessarily the case with the Asus board you tested, but maybe something we should be checking in future?

We? You have one post here, lol. The vast majority of the tests in our suite pass several minutes of full 100% CPU load. The Vegas Pro session, for example, runs roughly 15 minutes full-on stressed with a 8700(k). The theory of overheating VRM, throttling down and then kicking back in is, however, not just a plausible one, you could even consider it as normal behavior. We test with liquid cooling mostly, that will have a better effect on that processor area and VRM a bit as well. However, as far as B360 goes ... there are certain compromises made when opting the lowest budget motherboards with limited phases ... and then equip it with a high-frequency upper segmented proc. I'll monitor this a bit more in the future though, it's an interesting point.

Hilbert Hagedoorn:

We? You have one post here, lol.

I was speaking "we" as enthusiasts the world over in general. 🙂 I started looking into this a bit while planning an upgrade to i5-8x00. Tom's was the first to make me think about it in their review of the 8400 when they mentioned that its possible for a motherboard to shortchange a Coffee Lake in this way. That was theory though. Then I found TechSpots' finding. Granted that was with a really really cheap board and the most power hungry CPU, but it got me wondering if we might run into something similar with, say, a typical H370 board and a somewhat faster, but by no means "top" end, chip like the 8600. Food for thought.