Ryzen 7 node shrink

I was wanting to know the difference between a low power and high power node, And how different the clock speed would be giving the advancement from 12nm to 7nm. What would a realistic clock speed be for ryzen 2?

Attached: 0000999_amd-ryzen-7-1800x-am4-36ghz-8-core-processor.png (960x458, 298K)

Nobody knows user. Clock speed has as much to do with the design of the logic as with the semiconductor process. There are just too many factors to make a good guess.
BTW "7nm" is just a marketing name, there's no item on the die that's actually that size.

I guess they're not any good comparisons if you looked at Intel's node shrink......oh wait.

Attached: UqwMh6j1_400x400.jpg (400x400, 27K)

Honestly we may see a 300-500mhz improvement. If that is the case on top of IPC improvement Intel will be worthless.

Zen2 shouldn't be such a major departure in architecture from Zen+ and you could look at TSMC estimates for 7nm to make a rough estimate, no?

That's the thing, you can't compare both. They're totally different architectures made with different processes and optimized for different purposes (Intlel for ST throughput, AMD for power consumption).

Probably stock of 4.7 Ghz Turbo and maybe 5 Ghz overclock.

Either way, if AMD can get to 4.5+ ghz on stock, I'll get it.

I'm sure they have a pretty accurate assessment, I just don't know that they released it to the public yet.

Low power and high power nodes are usually differentiated by their transistor libraries and the voltages they're built to handle. High voltage transistors are physically different from low voltage ones.
This is pretty indepth material in solid state EE though.

Theres no way to predict what clockspeeds will materialize in 7nm Zen2 parts. Actual clocks in a final part may not even be what the foundry is capable of if they want balls to the wall. Economics of the matter still rule over everything.

the clock speed isn't really as relevant when downscaling transistors. if you can fit more transistors in the same space you can create more complex designs that can preform faster operations with the same or even lower clock speeds than previous models. while the clock speeds usually do increase at least a little, at this point the major advancements are being made in more efficient circuits. for example, floating point operations improved dramatically during the x32 years as transistors shrank and allowed for more complicated operations with more logic and less brute force. clock speeds are starting to level off but that doesn't mean processor advancements are going to slow down at all.

The word.

Actually there haven't been a lot of single thread throughput increases lately. It mostly just has been cheaper to put more cores into the chip because of more mature and thus more efficient processes.
And at some point in the near future transistors will be impossible to shrink down further. Then exotic tech like vacuum channel and graphene will allow big clock increases again.

i hope they do more research with proteins. i remember reading a few years ago about researchers trying to grow an organic computer, it sounded pretty wild. the idea was having a computer that could produce proteins where they were needed and slowly grow in processing power. it probably won't happen in my lifetime but i can still dream, lol.

The more important question is; will the techtubers start using amd processors for graphics card benchmarks if AMD is faster for gamer gaming gamers

Seems like the lifespan of something like that would be horrible.

the highest binned 12nm ryzen (2950x) already does 4.5 ghz

The 9900k already blows an overclocked 2700X in the water

Ryzen 3000 is most likely close +/- to the 9900k

I seriously doubt that, 8c 8t Intel vs 8c 16t 2700x would beat ryzen.

Nah, the 9700k is slower than the 2700X on every multithreaded benchmark out there

Attached: 1537707457895.png (1286x144, 13K)