Intel Skylake / Kaby Lake

[DrMrLordX]

What's with all this Ryzen garbage in the Intel thread?

Well . . .

Well it started with this quote "If the pretty strong rumors are true, Intel will be up against 16 core/ 32 thread Ryzen cpu's not AMD'S mainstream platform. AFAIK, Skylake-X will be offering a top end 12 core to compete with AMD's hedt."

Actually it started many pages ago with someone posting cherrypicked gaming benchmarks where Kabylake trashed Ryzen. Y'know, as if Intel had actually released anything new, which they hadn't . . .

 

[raghu78]

5.2GHz easily on 6 cores when KBL can't reliably do 5GHz OC with 4 cores?

https://newsroom.intel.com/newsroom.../11/2017/03/Ruth-Brain-2017-Manufacturing.pdf

see slide 21 for 14++ transistor performance.

14++ transistor performance is 12% higher than 14+ . That should be enough to reliably hit 5.2 Ghz . As for 6 cores at 5.2 Ghz the people who overclock have very good AIO or liquid cooling solution and power circuitry on high end Z370 motherboards will be more than enough to support stable OC >= 5.2 Ghz.

 

[formulav8]

14C, 16C and 18C beasts, if Intel eventually chooses to release them for desktop users.

You probably meant, if normal people could actually afford or would pay for that over-priced silicon. The Intel would gladly take your 5k for one of those cpu's.

 

[Arachnotronic]

5.2GHz easily on 6 cores when KBL can't reliably do 5GHz OC with 4 cores?

KBL can do 5GHz pretty commonly if you exclude AVX2 workloads. I have two 7700Ks purchased at retail and I run them both at 5GHz non-AVX2, 4.8GHz AVX2.

14nm++ should allow for >5GHz in non-AVX2 code pretty easily for decent samples, IMHO. But we shall see.

 

[itsmydamnation]

KBL can do 5GHz pretty commonly if you exclude AVX2 workloads. I have two 7700Ks purchased at retail and I run them both at 5GHz non-AVX2, 4.8GHz AVX2.

14nm++ should allow for >5GHz in non-AVX2 code pretty easily for decent samples, IMHO. But we shall see.

Its the extra 50% of active transistors that will be the challenge, 5ghz 7700k is what about 140watts with highend cooling, so @ 6 cores even with process improvement your going to be very close to 200 watts, if your running high end air its probably going to be even higher as your temps will be higher.

 

[Shivansps]

$400 for a 6c/12t CPU will loose to a Ryzen 8c/16t@$330

So, what is not true ?

And my "guesses" as based on facts about todays Intel and AMD chips.

You cant know that based on current Intel or AMD processors because Skylake-X doesnt exist yet, the only thing we can be sure off is that will be faster than client Skylake. Thats all.

You said Skylake-X had to face AMD HEDT, thats not true because AMD has a $500 mainstream chip, and the 1700X is too close to the $400 mark, maybe 10/12C will face AMD HEDT (and mind you, if 12C is MCC, there are probably higher core count ones). But 6C are gonna be inside AMD mainstream price range, period, i base this on the existence of the 6800K. And there is a chance that 8C will be as well.

Is also a fact that Ryzen cant match Broadwell-E (1st gen 14nm) core clocks, and client Skylake is already faster on IPC, is also a fact that Skylake-X also features a new cache structure, so im not talking just because the air is free here, the only thing it may not be is AVX-512, but i dont think they are gonna cut it.

 

[crashtech]

Knowing whether the upcoming Coffee Lake CPUs will be backwards compatible with at least the Z270 chipset would drive my decision making, which will be made in the next couple weeks. If such compatibility remains a rumor, it might be time for me to switch away from LGA1151.

 

[itsmydamnation]

You cant know that based on current Intel or AMD processors because Skylake-X doesnt exist yet, the only thing we can be sure off is that will be faster than client Skylake. Thats all.

Sure for workloads that can use avx-512 but outside that, Nope that is an assumption on your part. For example caches dont get bigger for free, you trade latency/power/area/size/associativity, end result is a different set of trade offs. What if skylake-X has higher IPC but then hits a lower Fmax because of reason X/Y/Z

 

[Markfw]

You cant know that based on current Intel or AMD processors because Skylake-X doesnt exist yet, the only thing we can be sure off is that will be faster than client Skylake. Thats all.

You said Skylake-X had to face AMD HEDT, thats not true because AMD has a $500 mainstream chip, and the 1700X is too close to the $400 mark, maybe 10/12C will face AMD HEDT (and mind you, if 12C is MCC, there are probably higher core count ones). But 6C are gonna be inside AMD mainstream price range, period, i base this on the existence of the 6800K. And there is a chance that 8C will be as well.

Is also a fact that Ryzen cant match Broadwell-E (1st gen 14nm) core clocks, and client Skylake is already faster on IPC, is also a fact that Skylake-X also features a new cache structure, so im not talking just because the air is free here, the only thing it may not be is AVX-512, but i dont think they are gonna cut it.

What ? would you stop putting words in my mouth ? I said that since a 14core 2.5 ghz Intel with a 35 meg cache (E5-2683v3) only bests my 1700x at stock by a few %, that a 6 core skylake-X I don't think could beat it. Nothing about HEDT. And 6 core 12 thread Ryzens are $219 ! So stop the FUD until you get something to prove your point.

And Ryzen can't beat Broadwell-E core clocks ???? Do they beat 3.7 GHZ stock ? I don't think so in 8 core or greater...The 6900k at turbo IS 3.7. Thats what my Ryzen does stock.

 

[happy medium]

but the Intel fanboys will not listen

wow , really?!

 

[IEC]

You guys expecting Coffee Lake (much less, Skylake-X) to be 5GHz are going to be disappointed. The 7700K was already outside of its ideal voltage to attain its stock clocks. Reaching 5 GHz on even 6 cores would be looking at 200W+ TDP...

 

[.vodka]

You guys expecting Coffee Lake (much less, Skylake-X) to be 5GHz are going to be disappointed. The 7700K was already outside of its ideal voltage to attain its stock clocks. Reaching 5 GHz on even 6 cores would be looking at 200W+ TDP...

It's possible...

I mean, it would be Intel's third attempt at client Skylake, plus a better process and physical implementation, plus the aid of delidding + CLU and a properly overengineered motherboard with a beefy VRM... Coffeelake's die size is going to be obviously bigger than Skylake/Kabylake, so that will help with heat transfer. Voltage required to get there will probably be similar to Kabylake, or a little better considering 14nm++ improvements.

If it doesn't hit 5GHz under these ideal (if not punishing considering the long term) conditions it'll be very close, or it'll be just like the 6700k with a little % actually getting there and not hitting a voltage/thermals/frequency wall based on die quality. We've had >200w CPUs before when overclocked, >4GHz Bloomfield (i7 9xx) was probably in that ballpark and it wasn't a problem getting that cool enough. You just had to deal with a PC doubling as a space heater


Skylake-X on the other hand... nah. No way it's going to overclock like Kabylake. Intel's server dies have historically overclocked worse than their mainstream dies, and this time you not only have a bigger L2 that could be a problem, but also the AVX-512 hardware that's even bigger and more power hungry than what's on client Skylake cores. Sure, you have 14nm+ as the base here helping... but stiIl, I'd be surprised if it overclocks worse than Broadwell-E. Probably similar top best case clocks (4.2-4.3GHz)


Having said that, it's going to be more of the same, and it's just Intel milking Skylake as much as they can until they get their next gen core out of the door in a few years or whatever. That 6 core Coffeelake is LONG overdue for the mainstream socket.

When was the last time Intel operated like this, staying so long on the same base architecture? The Netburst days with 4-6 years of polishing that turd through node shrinks/improvements and the natural single core -> dual core transition?

 

[ozzy702]

You guys expecting Coffee Lake (much less, Skylake-X) to be 5GHz are going to be disappointed. The 7700K was already outside of its ideal voltage to attain its stock clocks. Reaching 5 GHz on even 6 cores would be looking at 200W+ TDP...

I don't think most people talking about 5ghz 6-cores are talking about Skylake-X. On an improved process there is no reason to believe that 5ghz clocks won't be attainable on Coffee Lake. I'd be willing to bet that 4.8ghz is achievable with poor chips and 5ghz+ with winners of the silicon lottery. I doubt it would be 200w, let alone above.

 

[tamz_msc]

Improved drive current vs leakage current characteristics does not necessarily imply even higher clocks in the 5GHz territory, given that we don't know the operating voltages. People seem way overexcited from mere transistor performance curves and are already celebrating 5.2GHz for 6 cores when nothing is known for certain regarding voltages that might be required for such clock speeds.

6700K could reliably do 4.6-4.7 GHz, even 4.8GHz. 7700K reliably manages 4.8-4.9GHz. So that's a mere 200MHz considering 6700K@4.7GHz vs 7700K@4.9GHz; the latter comes at the cost of crazy temperature spikes and voltage requirements. But people still expect a 6 core CPU that is barely 1/6th bigger in size compared to its 4 core equivalent to hit 5GHz or more.

So much for rationality in some of the posts in this thread.

 

[frozentundra123456]

What ? would you stop putting words in my mouth ? I said that since a 14core 2.5 ghz Intel with a 35 meg cache (E5-2683v3) only bests my 1700x at stock by a few %, that a 6 core skylake-X I don't think could beat it. Nothing about HEDT. And 6 core 12 thread Ryzens are $219 ! So stop the FUD until you get something to prove your point.

And Ryzen can't beat Broadwell-E core clocks ???? Do they beat 3.7 GHZ stock ? I don't think so in 8 core or greater...The 6900k at turbo IS 3.7. Thats what my Ryzen does stock.

I would expect nearly everyone who has BW-E to overclock. I seriously doubt on average Ryzen overclocked can reach BW-E overclocked. I also expect Skylake-X will widen the clockspeed gap.

 

[frozentundra123456]

wow , really?!

Yea, really surprised at the tone of his posts lately.

 

[Markfw]

I would expect nearly everyone who has BW-E to overclock. I seriously doubt on average Ryzen overclocked can reach BW-E overclocked. I also expect Skylake-X will widen the clockspeed gap.

I doubt that. I would say more, but then I would have to infract myself.

 

[Markfw]

Yea, really surprised at the tone of his posts lately.

Well that has a lot to do with the fact that I OWN a couple of RYzens, and I am impressed with what it does compared to my 14 core Xeons.

But some people just conjecture without knowing any facts or READING about things.

 

[xdfg]

E5-2683v3: 14x 2.5 GHz = 35 GHz
R7 1700X: 8x 3.5 GHz = 28 GHz

Since these chips have the same performance, let's put to rest the myth of Intel IPC superiority. In fact, this is proof that Ryzen's IPC is 25% better than Haswell, since it's getting the same performance with only 80% of the clocks.

 

[.vodka]

E5-2683v3: 14x 2.5 GHz = 35 GHz
R7 1700X: 8x 3.5 GHz = 28 GHz

That's not how it works.

 

[xdfg]

That's not how it works.

If you have 3.5 billion clock ticks per second per core and 8 cores, how many clock ticks per second were there?
28 billion per second.

What is a hertz?
1/second.

What is 28 billion per second?
28 GHz.

What is IPC?
Instructions per clock tick.

If CPU A does the same work in 28 billion clock ticks that CPU B does in 35 billion clock ticks, which has more IPC?
CPU A

 

[crashtech]

Apps don't always scale linearly with clockspeed, and rarely scale linearly with core count, so the kind of simplistic calculations like cores times speed are ballpark guesstimates at best.

 

[inf64]

I think people find it hard to let go the preconceived notions they have regarding certain products. For example AMD has a stigma of having poor IPC (ie. Bulldozer) hence Ryzen is "slow". When in fact Ryzen is a masterpiece of perf./watt efficiency and a very high IPC uarchitecture that only lacks full AVX execution width- which currently means jack sh*t for average Joe consumer. Even in server space it (HPC) accounts for roughly 15% of total market. AMD is positioned really well, best in the last 10+ years to attack intel's x86 dominion. I have a good feeling about their x86 cores this time around .

 

[beginner99]

Yes, and those products won't be competing with Zen 2, but Zen 1. What value proposition does Raven Ridge offer (4C + Vega iGPU) compared to Coffee Lake-H or even Coffee Lake-U? Better graphics, but worse CPU? That combination hasn't exactly been a winner in recent years, so I'm skeptical.

CPU is almost on par now and laptop chips don^t clock to 4.5 ghz so intel doesn't have the clock advantage. Also looking at Ryzen the GF process seems actually pretty good at lower frequencies, better than Intel. If Vega also is efficient then we will finally get 4 core laptops that aren't huge 5 lbs beasts and actually last a day.

 

[beginner99]

$400 for a 6c/12t CPU will loose to a Ryzen 8c/16t@$330

In DC or encoding it will loose but maybe not in gaming due to clocks. However we don't know yet

I say this because gaming is more important for my home machine than doing "real work". For 6-cores the real competitor will be Coffeelake anyway unless they extra L2 cache of Skylake-x gives some magic gaming boost.