Intel Skylake / Kaby Lake

[beginner99]

If the new 6C CFL chip has a base below 3.5ghz, it will be stillborn. I won't even give it a second look.

Depends on price on OC capabilities. If it OCes like a 7700k and is priced like a 7700k for all I care it can have a 1 ghz base clock because I ain't running it at base clock.

 

[TheF34RChannel]

Depends on price on OC capabilities. If it OCes like a 7700k and is priced like a 7700k for all I care it can have a 1 ghz base clock because I ain't running it at base clock.

While agreeing with you I also want to be lazy this time around and just have a high boost clock and enable boost on all cores in the BIOS (need a boost of at least 4.5 to make me happy). I will need to lower the voltage probably.

 

[LTC8K6]

Depends on price on OC capabilities. If it OCes like a 7700k and is priced like a 7700k for all I care it can have a 1 ghz base clock because I ain't running it at base clock.

Well, let's look at a hypothetical 3.5ghz base clock.

To overclock like a 7700K it would need a 1.3Ghz overclock at a minimum.

That seems like pie in the sky.

There is always the wild card of the 14nm++ process, though. It could be a big improvement, or not.

 

[Timmah!]

I wonder, if we will see some leaks of the upcoming HCC Skylakes (7920x and up) in the coming days as a response to Threadripper reveal today.... to take some wind away out of AMD´s sails

 

[scannall]

I wonder, if we will see some leaks of the upcoming HCC Skylakes (7920x and up) in the coming days as a response to Threadripper reveal today.... to take some wind away out of AMD´s sails

It would probably be better if Intel dealt with their X299 problems. It has um, 'issues' still.

 

[Topweasel]

Well, let's look at a hypothetical 3.5ghz base clock.

To overclock like a 7700K it would need a 1.3Ghz overclock at a minimum.

That seems like pie in the sky.

There is always the wild card of the 14nm++ process, though. It could be a big improvement, or not.

I understand where you are coming from but it looks like 5.0+ and maybe even more is the capability of the process. With proper cooling and a delid at 10core SL-X is hitting 5.GHz assuming cooling and VRM support.

So realistically Intel isn't clocking the CL 6 core at 7700K speed because of power. Since its the same arch on the same process power requirements will be ~45-50% higher for the same clocks. But for overclocking I don't see why a 6 core die on this process wouldn't be able to hit 4.8-5 minus the mayo in between the chip and the heat spreader. It'll use more power and maybe some mid level boards won't work. But it doesn't seem like it will be an unreasonable OC.

 

[formulav8]

Skylake-SP cache latency: https://www.servethehome.com/wp-con...-SP-Microarchitecture-L2-L3-Cache-Latency.jpg

Thanks, thats what I've been wanting to know. Intel did well increasing the L2 cache so much with only a 2 cycle latency increase.

 

[LTC8K6]

I understand where you are coming from but it looks like 5.0+ and maybe even more is the capability of the process. With proper cooling and a delid at 10core SL-X is hitting 5.GHz assuming cooling and VRM support.

So realistically Intel isn't clocking the CL 6 core at 7700K speed because of power. Since its the same arch on the same process power requirements will be ~45-50% higher for the same clocks. But for overclocking I don't see why a 6 core die on this process wouldn't be able to hit 4.8-5 minus the mayo in between the chip and the heat spreader. It'll use more power and maybe some mid level boards won't work. But it doesn't seem like it will be an unreasonable OC.

The CFL 6 core is not on the same process as the 7700K though. 7700K is 14nm+ and CFL 6C is on 14nm++.

AFAIK, none of the other chips are on 14nm++? The CFL 6C will be the first we see of 14nm++.

I actually don't think it's an unreasonable overclock at all. I just think it's unreasonable from a low base clock. I don't expect a low base clock.

Theoretically, the CFL 6C chip should be much better than the 7700K in all ways.

 

[Topweasel]

The CFL 6 core is not on the same process as the 7700K though. 7700K is 14nm+ and CFL 6C is on 14nm++.

AFAIK, none of the other chips are on 14nm++? The CFL 6C will be the first we see of 14nm++.

I actually don't think it's an unreasonable overclock at all. I just think it's unreasonable from a low base clock. I don't expect a low base clock.

Theoretically, the CFL 6C chip should be much better than the 7700K in all ways.

I don't know enough about Intel's process's (or anybody's really) to be able to tell if they will get substantial improvement without a shrink. 14nm is 14nm to me. That's why I can see them bringing down the base clock. But inretro 3.6 does seem low.

 

[TheF34RChannel]

Still makes me wonder why that chip in the latest entry had such low clocks when the previous entry was at 4.2GHz... Any ideas? Could it be the non K part (still too low imo)? I can see the 4.2GHz being the actual finalised frequency but certainly nothing lower.

 

[Lodix]

14nm++ per Intel claims should bring the same increase in performance as 14nm+ from the original 14nm ( around 12% better performance in ideal conditions ).

 

[IEC]

Still makes me wonder why that chip in the latest entry had such low clocks when the previous entry was at 4.2GHz... Any ideas? Could it be the non K part (still too low imo)? I can see the 4.2GHz being the actual finalised frequency but certainly nothing lower.

ES chips almost always have different specifications than the final retail shipping product.

 

[TheF34RChannel]

ES chips almost always have different specifications than the final retail shipping product.

True, and in that I have invested my hope for a nicely clocked retail product.

 

[IntelUser2000]

14nm++ per Intel claims should bring the same increase in performance as 14nm+ from the original 14nm ( around 12% better performance in ideal conditions ).

Ideal, so it'll turn out to be far less. I wonder if even 100MHz increase at the same core is in order? Of course I'm talking about overclocking frequency. Overclocking headroom will decrease because base clocks might be upped more.

Ideal as in:
-What metric are they talking about? At 0.7v, which is practically idle?
-Probably when the clock isn't limited by being near the limits of frequency(~5GHz)?
-It doesn't have some quirks in the circuit that limits clocks

Its a real bad time to be in the high-end CPU design. Most of the consumer base heavily relies on clocks and perf/clocks, they don't scale with cores, yet that's the only thing you can do because you can't increase clocks anymore. Even if you could increase it massively most people don't need the extra performance.

 

[Lodix]

Ideal, so it'll turn out to be far less. I wonder if even 100MHz increase at the same core is in order? Of course I'm talking about overclocking frequency. Overclocking headroom will decrease because base clocks might be upped more.

Ideal as in:
-What metric are they talking about? At 0.7v, which is practically idle?
-Probably when the clock isn't limited by being near the limits of frequency(~5GHz)?
-It doesn't have some quirks in the circuit that limits clocks

Its a real bad time to be in the high-end CPU design. Most of the consumer base heavily relies on clocks and perf/clocks, they don't scale with cores, yet that's the only thing you can do because you can't increase clocks anymore. Even if you could increase it massively most people don't need the extra performance.

I have just started following the desktop CPU market so I don't know how kaby lake improved frequencies in overall range of products, but just looking at the i7/5 K version there is a 8% bump in turbo frequency. As you noted there is a limit around 5GHz so it will make a major difference in mobile and high core count products where they are lower clocked.

 

[LTC8K6]

For 4C/8T on 14nm++ if we just give the same clock boost that we got from 14nm to 14nm+:

SL 4.0B/4.2T
KL 4.2B/4.5T
CFL 4.4B/4.8T

I think it's safe to say that a 6C/12T CFL chip will not be near those speeds with the 2 extra cores.
But it doesn't seem out of the question to reach the SL speeds?

 

[Edrick]

To overclock like a 7700K

I think it is quite amusing that some people actually have that expectation of a 6 core CFL chip.

 

[Arachnotronic]

I think it is quite amusing that some people actually have that expectation of a 6 core CFL chip.

It's a pretty realistic expectation, TBH. A year's worth of circuit enhancements + new process should do wonders.

 

[Edrick]

It's a pretty realistic expectation, TBH. A year's worth of circuit enhancements + new process should do wonders.

How many times in the past has a 6 core CPU (from Intel) been able to OC as high as its 4 core cousin? Just using history as a guide. Now if CFL was going to be something like 10nm+, then perhaps I would agree. But this whole 14nm, 14nm+, and 14nm++ does not instill much confidence of 5Ghz on 6 cores. Now it is quite possible to get a few golden samples to reach that speed on custom water loops, but that may be it.

But that's just my opinion.

 

[LTC8K6]

How many times in the past has a 6 core CPU (from Intel) been able to OC as high as its 4 core cousin? Just using history as a guide. Now if CFL was going to be something like 10nm+, then perhaps I would agree. But this whole 14nm, 14nm+, and 14nm++ does not instill much confidence of 5Ghz on 6 cores. Now it is quite possible to get a few golden samples to reach that speed on custom water loops, but that may be it.

But that's just my opinion.

I don't think we have had this situation before.

This is a 6 core chip on a new process.

I'm not even sure it's correct to say it's just a new process.

How do you get a 6 core Kaby Lake chip anyway?

It must actually be an all new chip, I think.

 

[Lodix]

How many times in the past has a 6 core CPU (from Intel) been able to OC as high as its 4 core cousin? Just using history as a guide. Now if CFL was going to be something like 10nm+, then perhaps I would agree. But this whole 14nm, 14nm+, and 14nm++ does not instill much confidence of 5Ghz on 6 cores. Now it is quite possible to get a few golden samples to reach that speed on custom water loops, but that may be it.

But that's just my opinion.

If you only guide yourself by the name of processes and not actual info then you should not have any confidence at all in your arguments. Intel 10nm+ will have similar transistors perfomance to 14nm++.

 

[Zucker2k]

It's a pretty realistic expectation, TBH. A year's worth of circuit enhancements + new process should do wonders.

Process enhancements are certainly important, but Intel has to beat Ryzen convincingly as well as worry about an aggressive AMD near-future release. If SKL-X is any indication, Intel should clock these puppies a tad higher than usual.

 

[Lodix]

Intel needs an improvement in architecture and lower prices if they want to remain competitive against the AMD roadmap. When is the new arch supposed to be released ? I don't care about their confusing names.

 

[dullard]

SL 4.0B/4.2T
KL 4.2B/4.5T
CFL 4.4B/4.8T

I think it's safe to say that a 6C/12T CFL chip will not be near those speeds with the 2 extra cores. But it doesn't seem out of the question to reach the SL speeds?

Very crude math here:

I'll assume there are no hidden secrets and thus CFL is basically SL and KL on a more refined 14 nm process. I'll also assume that Intel is correct that the 14++ process has 23% more efficient transistors than the 14+ process. Finally, I'll also assume the rumors that CFL is 6C/12T with a 95W TDP (a slight increase over SL and KL) are correct. Based off of that information alone, and using the standard approximate power formula for a CPU core (P = f*C*V^2) you end up with this:

• A 6 core CFL with hyperthreading at 95 W could reach 3.98 GHz all core turbo at stock. Round that up to 4.0 GHz.

Unless Intel has an ace up their sleeve, then 6 core CFL all-core turbo will be just shy of the SL all-core turbo. I honestly think it'll be slightly less than 4.0 GHz (say 3.9 GHz) turbo unless Intel has something else that is not yet known.

 

[LTC8K6]

Very crude math here:

I'll assume there are no hidden secrets and thus CFL is basically SL and KL on a more refined 14 nm process. I'll also assume that Intel is correct that the 14++ process has 23% more efficient transistors than the 14+ process. Finally, I'll also assume the rumors that CFL is 6C/12T with a 95W TDP (a slight increase over SL and KL). Based off of that information alone, and using the standard approximate power formula for a CPU core (P = f*C*V^2) you end up with this:

A 6 core CFL with hyperthreading at 95 W could reach 3.98 GHz all core turbo at stock. Round that up to 4.0 GHz.

Unless Intel has an ace up their sleeve, then 6 core CFL all-core turbo will be just shy of the SL all-core turbo. I honestly think it'll be slightly less than 4.0 GHz (say 3.9 GHz) turbo unless Intel has something else that is not yet known.

Is the die bigger? Would that give us better cooling, or worse? Or is it a wash with more cores and a bigger die such that we can just leave it out of our guesses?
Intel could also solder the IHS?