130W on AMD vs 180W Intel. Nearly 75%.

Also, I'm worried that they're at "engineering sample, glad it didn't crash" stage. Optimistically this will hit shelves at Q3, and at this point we may see sunny-cove-based 10nm (Whole chip, not only the "chiplet") cpu from the blue team.

I think AMD's success will come down to pricing...

You are comparing power for the entire system. I think they subtracted 55W of idle power consumption from both to get only the cpu, then compare: 75/125 = 0.6.

But isn’t part of idle consumption due to the cpu?

Not very much. For example i9 uses under 13 watts at idle [1].

[1] https://www.tomshardware.com/reviews/intel-core-i9-7900x-sky...

It's just a guesstimate. From TFA:

>"If we take a look at our average system idle power in our own reviews which is around 55W, this would make the Intel CPU around 125W, whereas the AMD CPU would be around 75W"

I have no idea what percent of idle power consumption the cpu should be responsible for, from their treatment I would assume negligible.

Intel has been flashing the 10nm for 2+ years now. If anyone Intel should be worried if the can keep their schedule.

I wonder how AMD is planning long term. Zen did put them back on track, but I hope it's not a one shot era and that they will find ways to get strong for a decade or so.

Chiplets and power are the next dimension of competition. Yield, chiplets vastly improve yield. The limit as chiplet size drops, yield basically goes to 1. Erasure Coding for Silicon.

And then power, part of that is process and part is design. I am not skilled enough to call that, but as leakage current drops, then does power draw, static power draw, allowing for frequencies to scale over larger ranges. Hell, maybe went with some novel clockless design.

Either way, this is momentous.

I didn't see any news about leakage being resolved, do you have an article ?

clockless and low power.. exciting

It was supposition based on the power differences, I only do armchair solid state physics.

better than bedroom solid state physics I guess