14nm 6th Time Over: Intel Readies 10-core "Comet Lake" Die to Preempt "Zen 2" AM4

[IntelUser2000]

Poor Intel.. They did not deserved this.

We often talk like companies are individuals, when in actuality they consist of many.

For many underlings and managers that really tried, no they did not deserve this.
For those higher ups that deliberately chose to take this path, yes they do deserve it.

Not having a backplan has backfired. It may have cost several hundred million dollars for few backup plans. Nothing compared to what they will potentially lose. They are going to pay for the mistake, that's all.

The 95W 9900K still clocks at 4.2GHz all core so a 10 core chip will be able to fit under a 120W TDP under the same conditions.

The 9900K is a fake-95W chip that changes well-established TDP definitions just so it can be advertised as a 95W chip. Even sites like Anandtech are justifying Intel's decisions with their articles.

Intel used to have TDP that correlated well with power and thermals, and the definition was solid.

Starting with Kaby, or maybe Coffee, they changed the definition so it can pretend its only at 95W. They don't want you to know 6 or 8 cores using the same architecture and process uses much more power than the 4 core one. They are trying to cover their mistakes not being able to port to 10nm in a very desperate way.

Perhaps Coffeelake is called such because it runs as hot as a cup of coffee that was just brewed. Cometlake is even hotter, like a Comet that just fell!

 

[jpiniero]

10nm must be really underperforming for Intel to consider this route.

As I mentioned, clock speeds are the issue. But there's also the question of just how large the Icelake volume is going to be. Since mobile is all BGA, probally makes it easier since OEMs could just ship models with Comet or rebranded Whiskey alongside whatever they end up getting from Icelake within a single SKU.

 

[zinfamous]

This says the rumors have it using a dual ring bus / dual die.
Didn't we see some "5 core" leaks somewhere?



https://www.notebookcheck.net/Intel...0-core-Comet-Lake-S-desktop-CPU.368371.0.html

so Intel is going whole-hog into the "gluing-things-together" design and are now gluing their ring buses together?

 

[IntelUser2000]

so Intel is going whole-hog into the "gluing-things-together" design and are now gluing their ring buses together?

No, the initial leak was from a Chinese site and some mentioned about the dual ring so people who don't know Chinese said: "DUAL RING BUS OMGZZZ!", when the original poster said nothing about a dual ring.

Also the 5 core leak is about Lakefield, which is a Icelake/Tremont hybrid.

 

[TheGiant]

No, the initial leak was from a Chinese site and some mentioned about the dual ring so people who don't know Chinese said: "DUAL RING BUS OMGZZZ!", when the original poster said nothing about a dual ring.

Also the 5 core leak is about Lakefield, which is a Icelake/Tremont hybrid.

I like the modern news. Put a bulls..t into the papers, get points and then spend 100x the effort to get things right.

 

[LTC8K6]

so Intel is going whole-hog into the "gluing-things-together" design and are now gluing their ring buses together?

These days, with the media we have now, you'll probably be closer to the truth if you go opposite the story.

 

[jpiniero]

The 9900K is a fake-95W chip that changes well-established TDP definitions just so it can be advertised as a 95W chip. Even sites like Anandtech are justifying Intel's decisions with their articles.

Intel's chips have been violating the TDP for some time, after all that's how Turbo Boost 2.0 works. They're just being very generous with the stock max turbo clocks for review purposes even though it obviously wouldn't fit under the 95W TDP.

 

[coercitiv]

Intel's chips have been violating the TDP for some time, after all that's how Turbo Boost 2.0 works.

Turbo Boost 2.0 is not based on the idea of violating TDP, please stop spreading this around. The whole idea of PL2 is to take advantage of the thermal capacity of your cooling assembly, and use it as a short time interval of elevated power. Once that capacity is depleted (meaning your heatsink heats up and reaches optimum heat dissipation an) the CPU goes down to PL1 and stays there indefinitely or until a low power usage interval allows for the cycle to begin again.

What Intel and mobo makers have been doing lately, and culminated with 9900K, was to allow this PL2 limit to make the CPU work completely out of stock spec: they took a +30% increase from PL1 to PL2 and a time limit of 30-100 seconds and transformed it into a +100% power increase with a virtually unlimited time period. This is not how Turbo Boost is supposed to work, it's a complete mess that sooner or later will come back and bite them where the visible spectrum cannot reach.

 

[Topweasel]

Intel's chips have been violating the TDP for some time, after all that's how Turbo Boost 2.0 works. They're just being very generous with the stock max turbo clocks for review purposes even though it obviously wouldn't fit under the 95W TDP.

Yeah but its been getting more and more egregious. AMD's moved into Intel's point of slightly understated TDP with most suggested performance coming from within the rated Power limit. But Intel these last 2 gens TDP has become barely the CPU's power usage a minimum rated performance with anything above that being much much more power usage.

 

[IntelUser2000]

Intel's chips have been violating the TDP for some time, after all that's how Turbo Boost 2.0 works.

You have it wrong. Turbo 2.0 goes over PL1(that's TDP) but it has to average down. The over-TDP part is called PL2. They changed the definition recently, and quietly.

Think about it. If its a 15W TDP rated chip, then the designers must design its cooling system to accommodate that. Hopefully you can make a thermal solution with a little bit of headroom. Also the TDP part is important because it guarantees over a longer period of time(remember average?) it uses no more than 15W. 15W used = 15W needed to cool. Modern CPUs with advanced power management systems can easily do this.

I like the modern news. Put a bulls..t into the papers, get points and then spend 100x the effort to get things right.

Yea, and the thumbs up/favorite/heart/like system I am not a fan of. You could get the worst post that's totally inaccurate but it'll have thousands of likes. Masses are not the best judges for accurate content. It's all driven by emotion.

 

[Zucker2k]

Intel's chips have been violating the TDP for some time, after all that's how Turbo Boost 2.0 works. They're just being very generous with the stock max turbo clocks for review purposes even though it obviously wouldn't fit under the 95W TDP.

Why would it fit under the 95W TDP? It's Turboing. That's what you get in the real world. If you don't like the extra performance put a small heatsink on it. That'll bring it to it's knees really quickly.

 

[IntelUser2000]

Why would it fit under the 95W TDP? It's Turboing.

Ok, its not a big deal for enthusiast desktops. They'll ignore the stock heatsink and buy one that has 3 fans and weighs more than a kilogram. Probably.

But it matters everywhere else. And the original definition of Turbo 2.0(introduced with Sandy Bridge) states Turbo takes advantage of thermal headroom when the heatsink is initially cool. If you haven't used the computer in a while, or you are doing something bursty(like web browsing, or using it intermittently) the heatsink will cool down.

Power usage goes up instantly when a CPU is loaded, but the heatsink will take seconds to minutes to fully heat up and reach an equilibrium. That's what Turbo 2.0 does. It uses the "seconds to minutes" difference to Turbo as much as it can, giving you extra performance.

The way they are doing it with 6, 8, 10 core CPUs are merely violating TDP specs.

 

[jpiniero]

What Intel and mobo makers have been doing lately, and culminated with 9900K, was to allow this PL2 limit to make the CPU work completely out of stock spec: they took a +30% increase from PL1 to PL2 and a time limit of 30-100 seconds and transformed it into a +100% power increase with a virtually unlimited time period.

OEMs/Board makers have presumably always been able to manipulate what PL2's TDP was and how long, but they were still limited by the max ratio if locked. Intel could have kept the max FCT to like 4.2 or 4.3 instead of the 4.7 it is.

 

[IntelUser2000]

OEMs/Board makers have presumably always been able to manipulate what PL2's TDP was and how long,

Some flexibility was likely always offered, but as coercitiv states, now they are doing it to pretend the new chips are "95W". It's not 95W. It's anything but. Maybe the 4 core Core i3 chips are 95W.

 

[Topweasel]

Ok, its not a big deal for enthusiast desktops. They'll ignore the stock heatsink and buy one that has 3 fans and weighs more than a kilogram. Probably.

But it matters everywhere else. And the original definition of Turbo 2.0(introduced with Sandy Bridge) states Turbo takes advantage of thermal headroom when the heatsink is initially cool. If you haven't used the computer in a while, or you are doing something bursty(like web browsing, or using it intermittently) the heatsink will cool down.

Power usage goes up instantly when a CPU is loaded, but the heatsink will take seconds to minutes to fully heat up and reach an equilibrium. That's what Turbo 2.0 does. It uses the "seconds to minutes" difference to Turbo as much as it can, giving you extra performance.

The way they are doing it with 6, 8, 10 core CPUs are merely violating TDP specs.

It can be enthusiasts caught out on this as well. I mentioned in the past that I have built up my last two systems to be on the quiet side. Big cases, Big fans, big coolers. With the idea of keeping the noise level down. Well fast forward to Intel's new TDP instead of being in the normal realm of slightly above actual power usage for a moment or two a 65w CPU under any real load is going to run at nearly 90w and 90-100w CPU at 120-130w. So now all that headroom that I designed to keep my system quiet is going to run it faster. Not an end of the world issue, but I shouldn't have to dig through dozens of options to get my CPU to stay with in spec just because keep my system cool. If I wanted a CPU to run out spec that much I would be an active overclocker again but I haven't done that since I got my X2 4400+ back in the day.

 

[jpiniero]

Some flexibility was likely always offered, but as coercitiv states, now they are doing it to pretend the new chips are "95W". It's not 95W. It's anything but. Maybe the 4 core Core i3 chips are 95W.

The TDP advertised has always been measured at base clock.

 

[IntelUser2000]

The TDP advertised has always been measured at base clock.

Base is a guarantee. However depending on your workload, the CPU may or may not reach TDP, because power used depends on how loaded the execution units and its resources are. That's why AVX Turbos are lower.

Enter Turbo. Turbo takes advantage of the headroom offered in such scenarios when the base clock can't reach TDP.

If I wanted a CPU to run out spec that much I would be an active overclocker again but I haven't done that since I got my X2 4400+ back in the day.

Yea, and they are basically selling you factory overclocked chips, because at 5GHz, it has no headroom. Sandy Bridge 2600K was clocked at 3.4GHz base, and 3.8GHz max Turbo and you could reach 4.4-4.5GHz with all cores meaning great OC headroom.

9900K has almost no headroom

 

[Topweasel]

The TDP advertised has always been measured at base clock.

You say that but look at the 7700k review on this site Versus the 9900k. The 7700k is almost exactly on target as a 90w CPU, the review for the 9900k it's a 95w CPU running at 160w at full load. 66% more power usage. The review for the 9900k mentions that the 6700k might run 20w higher but their review for that doesn't show it and I believe that it was mostly an AVX issue. But 25% while bad still isn't 66% more and noted here it was a peak at the time and not sustained.

 

[jpiniero]

You say that but look at the 7700k review on this site Versus the 9900k. The 7700k is almost exactly on target as a 90w CPU, the review for the 9900k it's a 95w CPU running at 160w at full load.

That's because the board is configured for PL2 = 210 W, time = unlimited.

 

[Topweasel]

That's because the board is configured for PL2 = 210 W, time = unlimited.

I am not question why its clocking that high. I understand that. Just saying that its a major switch even from Intel on how TDP relates to their speeds and turbo's. Pretty much destroying TDP as a spec.

 

[PotatoWithEarsOnSide]

I'm not sure that they did rate it any differently from how the 7700k was rated. If anything, they gave a conservative TDP rating for the 7700k; the generally accepted next band is 65w, and the 7700k was likely above that.
The reason why power draw was near the rated TDP limit is likely similar to why the 9900k was hitting 150w; the mobo manufacturers fiddling with p-states. The only reason we notice it now is because the 9900k operates outside of TDP when defaulted to out of spec, whereas the 7700k was still within TDP limit when defaulted out of spec.
What has changed is our actual understanding of how Intel lists its TDPs, and how Turbo Boost is actually operating.
Contrary to my initial opinions of the 9900k, it does appear to operate to spec within the TDP limits when the mobo manufacturers use Intel's recommended spec. The 9900k is only a space heater because the mobo manufacturers are "cheating" the specs. Clearly, it favours Intel not to prevent this, but their claims are technically correct.
What I'd be interested in is a list of benchmarks that are so short to perform that their results are entirely the product of Turbo Boost, and not remotely related to sustainable load; a 9900k at 4.2GHz versus the 4.1GHz you get with a 2700x would be no ce to see as a "stock" comparison.
TBH, I'm beginning to question relative IPC claims given the lack of understanding as to how processors were operating.

 

[IntelUser2000]

Clearly, it favours Intel not to prevent this, but their claims are technically correct.

Intel has so much power, that they are just as guilty standing around doing nothing.

Think of it how it works in legal terms. If you stand by watching and doing nothing, you are guilty on some level too.

Of course its worse in this case, because Intel is directly responsible.

https://www.anandtech.com/show/13400/intel-9th-gen-core-i9-9900k-i7-9700k-i5-9600k-review/21

The worst thing is they work with reviewers to promote this new TDP nonsense. Is Anandtech saying "maximum CPU power draw" for a number that's shown for few seconds, maybe for a minute? Nope, its for sustained numbers.

In this case, for the new 9th Generation Core processors, Intel has set the PL2 value to 210W. This is essentially the power required to hit the peak turbo on all cores, such as 4.7 GHz on the eight-core Core i9-9900K. So users can completely forget the 95W TDP when it comes to cooling. If a user wants those peak frequencies, it’s time to invest in something capable and serious.

This is stupid. PL2 has to come down after maybe a minute so it adheres to PL1, which is TDP.

It conveniently fails to mention that fact. If you pulled that crap for mobile, vendors would laugh at your face. Blame the motherboard manufacturers, my ass. What does the author of the review recommend? Let me repeat the quote said by another site which I won't mention.

Just buy it(and invest in a hefty cooling solution).

 

[Topweasel]

In this case, for the new 9th Generation Core processors, Intel has set the PL2 value to 210W. This is essentially the power required to hit the peak turbo on all cores, such as 4.7 GHz on the eight-core Core i9-9900K. So users can completely forget the 95W TDP when it comes to cooling. If a user wants those peak frequencies, it’s time to invest in something capable and serious.

That really is the rub there. Intel actively sells this as 95w CPU, but they also sell it as a up to 5GHz CPU per their page. Sure they also give you base performance. But they don't say "hey anything above base performance uses more power than our rating (which hasn't been the norm till now)". But in fact to get this CPU to running ad advertised (which is not base speed) you need twice the cooling power. This is fine print crap we should be past and well if it has to be there, you know, actually put it in the fine print.

https://ark.intel.com/products/186605/Intel-Core-i9-9900K-Processor-16M-Cache-up-to-5-00-GHz-

Without a single note on the site of max power draw. Unless I was digging through the forum, how am I supposed to figure out what size cooler to get. 95w+ like the TDP implies. Nope that's out. How about a more run of the mill enthusiast 120w cooler? Nope no good either. Well lots of LGA2011 CPU's ran at 140w would one of those work? Nope. Now we are getting into dual chip TR or TR2 territory with ~160w cooler required, but not sure, and probably sometimes more because not all silicon is going be the same. This isn't a big issue if you know ahead of time, but now that's on us to make sure people know because Intel sure doesn't. Then the final nail in this power coffin is that this 90w CPU no matter what level you are looking at it requires pretty much a cooler rated at twice its rated power usage to not be running at it's thermal limit. It's not great science to say this but whether its a CPU or Powersupply or Video card I don't like them sitting at their limit for extended periods of time and treat it as running your car at the rev limit for a long time. Now we are talking about any load whether you got a 90w or 160w cooler the CPU is running at it's cutoff temp. For me if I got this CPU I wouldn't feel comfortable without something like a 250w cooler to keep the CPU far enough off it's limit.

There is just so much bad about the power rating on this CPU. The problem is they wanted to keep it accessible to OEM's but the fact is if you say your CPU is 160w CPU it isn't the end of the world. I know I wouldn't be nearly as bothered by buying a known 160w CPU then finding out I am not seeing nearly the upper end clocks because I mistakenly thought a 120w cooler would be more than enough headroom for a 95w CPU.

 

[jpiniero]

Intel has actually required a 130W cooler for K for some time.

 

[Topweasel]

So they require a 130w cooler for a 90w CPU?