hmm, I think ring bus bottleneck should be something that needs a bit more quantitative data on, not just some theoretical overview or tech demo article. If mesh is so superior to RB for anything above 6 cores then surely the 9900K should be based on mesh right? but all the leak till now pointing toward 9900K is a 2-more-cores-added-RB-8700K if that the case then surely 9900K RB should be slower than 7820X when operating at the same clock, which I pretty much doubt it. Skylake HCC with mesh seems like an experiment just like adding FIVR into Haswell and later remove it. Mesh has its merit but the crucial thing is when.