Why do Intel CPUs generally support higher clocks than AMD?

Why do Intel CPUs generally support higher clocks than AMD?
What's their magic sauce?

Attached: 1137px-Intel_80386_SX_die.jpg (1137x1024, 455K)

process advantage on 14nm. not really magical.

Speculative computing?

Clock speed doesn't really matter. It's the other instructions under the hood that make these processors so fast. The general x86 architecture is just what gets exposed to the user and the developer. The model architecture like somethinglake or shittytrail or roozen or whatever has additional hardware that works internally which isn't exposed to the user many times.

This is sort of what I'm talking about.

Attached: afe66e497e7ec52a3e6c841b635e34eb9790342c40fa8d9d1cdcd717f4fa8ec0.png (645x614, 83K)

That doesn't affect clock speed brainlet.

This. Any kind of logic gate level arrangement can speed up the execution of your software thread, but doesn't affect the speed the CPU can be clocked at (for the most part I guess).

That's why I put the question mark there. I wasn't really sure.

Intel managed to extract more juice from x86 by ignoring many security steps hoping no one would notice.

Yeah, but Intel *is* faster in clock speed 99% of the time too. And there's no reason to think AMD doesn't also have those optimizations, and they are slower just because they have slower transistors.

Attached: joker this.gif (300x186, 446K)

>they have slower transistors

Attached: 533.jpg (900x729, 67K)

Unsecure cpus.

Oh, you come from /v/? That's cute.

It's literally true though. Faster swtiching transistors -> higher clock speeds.

I'm a PC brainlet and have been looking for a CPU upgrade from my i5 6500 to something better, motherboard won't support 8th gen intel so I'm stuck on 7th, what's the best 7th gen i7 I can get that'll last me awhile?

Attached: 05784252.png (250x250, 79K)

The silicon jew

>what's the best 7th gen i7 I can get that'll last me awhile?
none.
you are stuck with i-corelets, that are expensive and difficult to cool.
stay where you are and in 6-9 months bite the bullet and get a ryzen at 7nm.

>Strict design rules
>optimize by hand like Apple's Ax chip
>200GHz finfet

Attached: 1537082296855.jpg (376x376, 15K)

Right. And what is their advantage?

Few more years of refinement

AMD bet all their money on server success and manufactored all their Ryzen chiplets on a power efficient process which allows them to make a 32 core monstrocity with the same TDP their old Vishera had. That process consumes a lot less power but doesn't go as fast.

That's why.

So do AMD consumer chips actually dissipate less power than their Intel equivalents?

intentional misinterpretations make you look like a newfag or retarded, which in this case i guess you're both.

AMD 8 cores disipate about the same power as Intel 4 cores.

At what clock speed?

t. assblasted /v/ermin retard

AMD used GloFo's 14 nm low power process. it's optimized for low power (hence the name) but sacrifices stability at higher frequencies. can't hold all those cores at 5 giggahurtz without ludicrous cooling.

at their respective stock speeds.

There's no 8 core Coffeelake to compare to the 2700X at the moment. Only way to compare it core for core/clock for clock.

Its Samsung's 14nm LPP. The moniker lower power plus is just marketing, the reason why it got that designation was for its light source, a high intensity laser produced plasma.
It was however designed primarily for low voltage moderately clocked ARM SoCs. GloFo had to tweak a vt for AMD to push clocks a bit higher. The issue with the process is voltage scaling.

Intel isn't a lightweight when it comes to their foundry business, even if they are completely face planting. 32nm, 22m, Trigate, even the 14nm process have absurd scalability. They have a few specialized vts, but they can still scale ridiculous well from 2ghz to 5ghz.

Oh, so AMD is piggybacking on the back of Global Foundries while Intel has its own process?

from about 45nm until 14nm they had the process advantage

also deeper pipeline requires faster clock speeds (as seen in the pentium 4) since any cache misses are very costly

no but the reverse is true. if you want speculative execution to be effective, you need higher clock speeds. deep pipeline as stated above.

Attached: computer-system-architecture-lecture-note-6-hardware-performance-8-638.jpg (638x479, 63K)

Huh? How would a slower clock increase cache misses or the cost of cache misses? If anything it would decrease the cost of cache misses relatively speaking because the memory controller would be at less of a disadvantage compared to the CPU.
Not saying you are wrong I just don't understand why you say that.

remember AMD sold off their foundries when they got in trouble after Intel took the lead with core.

a faster clock speed compensates for cache misses

t. brainlet parrot

Oh yeah I can see what you mean now. Both having a faster clock and a deeper pipeline require having a cache with speculative execution to be worthwhile because if not then the RAM becomes a huge bottleneck.

because Intel takes so many liberties with safety of the processors that script kiddies now can easily fuck your shit up

There are also a few tricks. For instance if you improve the speculative execution or avoid draining the pipeline or improve cache hits you will improve overall performance.

yes and that's exactly the reason intel fucked itself in 10 nm process.

Jesus christ, prescott

Attached: 1476643449642.png (801x1500, 973K)

AMD FX-9590 was the first CPU to have 5GHz turbo without overclocking. Clock speed isn't worth much without good IPC - instructions per clock. That is why the FX-9590 was outperformed in single threaded tasks by competing Intel CPUs with lower clock speeds.
Today Intel leads the clock speed numbers, but not so much in IPC. Zen cores are cheaper to produce, cheaper to assemble into a CPU, and are power efficient. That is why any Intel CPU is answered by a more powerful AMD CPU which has more cores.
AMD reaching 7nm process while Intel being stuck at 14nm will mean that for the first time in ages, AMD will also be better for single threaded tasks. They might still not be able to reach 5GHz again, but with greatly increased IPC, a 4.5GHz AMD chip on the 7nm process will outperform a 5.3GHz Intel chip, even in games.
Intel will have an answer eventually, but until that happens they will lose a lot of ground to AMD in non-laptop markets (servers, databases, professionals, gamers, budget PC, and maybe even Apple ditching Intel and bringing their cult to the red side).

Attached: 20180526_141725.jpg (439x322, 93K)