Intel Corp (INTC)

[Unkwn]

In the Ryzen vs Skylake comparison, yes the core is bigger on the latter, but so is the performance. Skylake versus Ryzen core difference is much smaller if you compare it to the consumer Skylake which lacks AVX-512 and the extra L2 cache.

On the consumer side, Kaby and Coffee can clock 5GHz or more, while Ryzen struggles beyond 4GHz. That's in no insignificant way helped by the better performing transistors. Higher performing circuits also cost more in terms of die area.

Yes, that's true and I did acknowledge that. In the end, what counts, is performance vs. area. I posted the L3 cache area and AMD realizes a significantly smaller area with the same cache size. I don't have clock rates for either CPU for the L3 cache, maybe someone can share those. I have a lot of doubt that the difference is merely due to the higher clock rate of Intel's CPU. Even if it would be, it would be to question if that was a smart design decision, since it raises costs for each CPU significantly. It would be interesting to have equal numbers for lower clocking parts, e.g. the mobile ones, though those are more difficult to compare due to integrated graphics.

Anyway, even taking that slight performance advantage from Intel into account, I see no correlation to what Intel claimed in its slides of having as density advantage over the foundries. It's nothing more than hot air.

And yes, I fully acknowledge that Intel has a clock speed advantage with its own process. They are tweaking their processes towards clock speed whereas the foundries are more interested in density for most applications. We'll see what they can do for AMD at smaller nodes but it certainly isn't their main focus area. That remains a plus for Intel's custom processes. The big question is, whether those few percent better single thread performance justify the higher costs for the larger chips. Time will tell.