Discussion Intel current and future Lakes & Rapids thread

[dmens]

I can't believe I have to explain this, but overclocking, by definition, is going beyond what the manufacturer offers out of the box.

Now that's enough feeding the trolls from me. Anyone with eyes can look at our two predictions and see who was right.

Bunch of dilettante nonsense. That is like saying if you ignore the TDP and turn the voltage to 11 and the chip gets to some frequency it is overclocking, but if the manufacturer does the exact same thing it is stock.

Either way, you have no point, because 257 package watts and still no 4ghz atom. Figure every core in there is already rocking 1.4V. inb4 1.4V is stock voltage LOL.

 

[biostud]

With the current GPU market, Intel could release an Etch-a+Sketch and still sell every one they make for several hundred over MSRP!

probably, but for me it does not make any sense to upgrade my CPU until I can actually buy a video card as well…

 

[itsmydamnation]

I can't believe I have to explain this, but overclocking, by definition, is going beyond what the manufacturer offers out of the box.

Now that's enough feeding the trolls from me. Anyone with eyes can look at our two predictions and see who was right.

Honestly to everyone else you look like an idiot...... just saying.

Here's a question for you if intel released a ~224 core ATOM 4 tile SPR type chip what would the voltage , clocks and power per core be.

All you need to do is look at the core design.. 3 cycle L1D..... lets look at all the other cores on modern processes with 3 cycle L1D , lets look at their voltage , clock and power consumption. but atom is special, they can bend space and time .... am i right ?

 

[Zucker2k]

You do realize the high PL settings are overclocking, right? Or do you actually think maximum thermal velocity boost/PBO/whatever is "stock"? Don't answer, that is a rhetorical question.

Try actually look at the voltage limited Tremonts versus blowout voltage Gracemonts to get to this frequency before claiming any sort of "scaling". Pro-tip: shoving more volts to get more frequency is not scaling, it is desperation.

And what figures are you basing all these on? Your desperation borders on grudge against your former employee to the extent that you have to pull numbers from thin air to bash the work of your betters. That's just pathetic.

 

[dmens]

And what figures are you basing all these on?

Haha, it is called engineering reality. Go look at the Tremont reviews and datasheets. Tremont-based products have a strict thermal envelope, Alderlake obviously does not. It is not that hard to understand.

your betters

Francois? Is that you?

 

[itsmydamnation]

And what figures are you basing all these on? Your desperation borders on grudge against your former employee to the extent that you have to pull numbers from thin air to bash the work of your betters. That's just pathetic.

So if we draw a generic perf / power curve for 3.7ghz gracemont where on that curve do you think you will find, i bet you will find it being nothing more then a guard band from the very end of the curve.

The funny thing is i dont think anyone has a bad thing to say about the new atom cores, whats stupid is the ignoring of architecture , the ignoring of relative positioning on the power/perf curve to try and justify the stupid linear extrapolation people have done. Sure its got near the linear extrapolation point by doing the equivalent of moving from a balanced position in the curve to the end of the curve.

Will you be happy if we start comparing LN2 or phase change runs of Zen4 to air cooled ADL, because in terms of power/perf curve moment thats about the same as the 3.3ghz ~2 watt a core tremont to 3.9ghz gracemont.

 

[jpiniero]

Haha, it is called engineering reality. Go look at the Tremont reviews and datasheets. Tremont-based products have a strict thermal envelope, Alderlake obviously does not. It is not that hard to understand.

It's a desktop K product, what do you expect. They are definitely pushing the power draw but that was to be expected.

 

[Exist50]

Honestly to everyone else you look like an idiot...... just saying.

So I look like an idiot for correctly predicting/extrapolating Gracemont frequencies? I must say, that's a bold strategy.

 

[dmens]

So I look like an idiot for correctly predicting/extrapolating Gracemont frequencies? I must say, that's a bold strategy.

Hah, no 4ghz in sight. And wasn't your extrapolation literally a straight line you drew from iso-power Icelake to Tigerlake frequency increase due to magical superfin scaling, therefore according to you, Gracemont ought to be able to hit 4ghz iso-power w.r.t Tremont max boost? Is that the prediction/extrapolation you are still holding onto? That is not looking so good for you.

 

[majord]

Doesn't really matter. If it does boost to 4Ghz , but runs a much higher VID to get there you can both have a prize.

Either way it's a moot point. 250w for year+ old 5950x performance is disaterous perf/watt. Looking like These e cores are , as predicted a waste of time on the desktop, merely a failed attempt to compensate for GC's poor area and power efficiency

 

[Hulk]

Doesn't really matter. If it does boost to 4Ghz , but runs a much higher VID to get there you can both have a prize.

Either way it's a moot point. 250w for year+ old 5950x performance is disaterous perf/watt. Looking like These e cores are , as predicted a waste of time on the desktop, merely a failed attempt to compensate for GC's poor area and power efficiency

Good point but on one hand you are writing that desktop power is a disaster and then in the next sentence you're implying desktop power is irrelevant (e cores a waste of time on desktop)? Perhaps these ADL systems might run at very low power on the e-cores much of the time and only ramp up the watts for shorts bursts for many user scenarios. Make actual real-world power usage not so high. Performance leaks suggest ADL is going to be competitive with the 5950X in some areas and may cost less. I think ADL is going to be much better than RKL. Where RKL was basically behind Zen 3 everywhere there may be usage cases where ADL might be a better solution for Zen 3. Don't get me wrong Zen 3 was and still is a masterpiece, but we should give Intel some credit for having the guts to bring something completely different to the arena.

 

[Exist50]

Hah, no 4ghz in sight. And wasn't your extrapolation literally a straight line you drew from iso-power Icelake to Tigerlake frequency increase due to magical superfin scaling, therefore according to you, Gracemont ought to be able to hit 4ghz iso-power w.r.t Tremont max boost? Is that the prediction/extrapolation you are still holding onto? That is not looking so good for you.

I actually said 3.9 was in line with that, and I expect 4GHz to be hit with an overclock. But trolls lying to deflect from their own embarrassment is nothing new.

 

[dmens]

I actually said 3.9 was in line with that, and I expect 4GHz to be hit with an overclock. But trolls lying to deflect from their own embarrassment is nothing new.

So to summarize, you missed both 3.9/4.0ghz marks even with unlimited power feeding those cores, and your extrapolation is an iso-power straight line so your technical justification is also wrong.

Man, you are a disaster.

 

[eek2121]

Bunch of dilettante nonsense. That is like saying if you ignore the TDP and turn the voltage to 11 and the chip gets to some frequency it is overclocking, but if the manufacturer does the exact same thing it is stock.

Either way, you have no point, because 257 package watts and still no 4ghz atom. Figure every core in there is already rocking 1.4V. inb4 1.4V is stock voltage LOL.

Haha, it is called engineering reality. Go look at the Tremont reviews and datasheets. Tremont-based products have a strict thermal envelope, Alderlake obviously does not. It is not that hard to understand.



Francois? Is that you?

So to summarize, you missed both 3.9/4.0ghz marks even with unlimited power feeding those cores, and your extrapolation is an iso-power straight line so your technical justification is also wrong.

Man, you are a disaster.

Really? Guys? Are you serious? Nobody here knows the amount of power that Gracemont consumes. I'm willing to bet that Golden Cove is 75+% 257W of power draw. Look, I get that you guys might like one company more than another, but it's time to face the music, Intel likely has made an awesome chip. Under a power virus scenario, a QS rated chip with a modified PL2 is consuming 257W of power. If that chip were stock it would not be consuming 250+W of power. Sure, Intel is going to be lax about PL1/PL2 again, leading to more abuse by motherboard manufacturers, but this chip ain't rocket lake. Also, you are looking at the flagship parts, the real parts to look at are midrange and mobile.

BTW, I could make my 5950X consume 250+W of power if I wanted, I just choose not to. Let's wait for reviews/benchmarks.

 

[dmens]

BTW, I could make my 5950X consume 250+W of power if I wanted, I just choose not to. Let's wait for reviews/benchmarks.

AMD could have set the maximum power to 250W, but chose not to because they achieved total victory on perf/W. But please, crank that 5950x to 250W and post benchmark results. It is always fun to see where things really stand, again. We already know that would be a massacre even against ADL.

 

[Hulk]

Really? Guys? Are you serious? Nobody here knows the amount of power that Gracemont consumes. I'm willing to bet that Golden Cove is 75+% 257W of power draw. Look, I get that you guys might like one company more than another, but it's time to face the music, Intel likely has made an awesome chip. Under a power virus scenario, a QS rated chip with a modified PL2 is consuming 257W of power. If that chip were stock it would not be consuming 250+W of power. Sure, Intel is going to be lax about PL1/PL2 again, leading to more abuse by motherboard manufacturers, but this chip ain't rocket lake. Also, you are looking at the flagship parts, the real parts to look at are midrange and mobile.

BTW, I could make my 5950X consume 250+W of power if I wanted, I just choose not to. Let's wait for reviews/benchmarks.

This seems reasonable to me. First of all we still have A LOT of unknowns. Second, yes, Intel needed to push clocks as high as possible in order to try and be competitive with AMD. Up to this point in time AMD has the better process and architecture, by a long shot with Zen 3.

But if early indications are accurate ADL and ESF may be closing that gap considerably and in some usage cases reversing it.

Intel's parts seem to take voltage/power well and clock very high with enough power and cooling. It is possible to push a 5800X to 5GHz all core under load without extreme cooling? I'm not saying AMD has needed to do this, I'm just wondering if it is possible? You can bet that if Intel does indeed catch AMD then they (AMD) will start pushing clocks and thermals if they have the leeway to do so.

Zen 3 is a proven winner, no doubt about it. Let's see how ADL stacks up.

 

[Shmee]

Lets remember not to throw out insults and go back and forth bickering with each other. Please keep the discussion civil and avoid personal attacks, stick to the tech. Thanks. -Moderator Shmee

 

[IntelUser2000]

At one point we were happy that Alderlake top part might perform like a 5900X.

Why is it so bad that it might perform like a 5950X? Even if it requires 250W to do so, that's a huge jump in one generation.

When it comes to arguments, isn't that called moving goalposts?

It doesn't even look like Alderlake was aiming at desktops at all. A mobile only chip being competitive with the 5950X sounds pretty good to me. How will it do on mobile I wonder hmm?

Oh, and I won't be surprised if the Gracemont-based successor to Jasper Lake has the same 3.9GHz ST, 3.7GHz MT frequency at 10W. It won't be able to maintain 3.7GHz MT in all scenarios, probably only in certain bursty workloads. But that's the whole point of Turbo. Maximizing performance up to the limit of thermals/power.

Every generation since Goldmont it was ~30% perf/clock gain plus whatever clocks needed to get an overall gain of 50%. Goldmont Plus is an overall 50% faster than Goldmont. Tremont is an overall 50% faster than Goldmont Plus. Gracemont will probably do similar.

Have you guys also forgot one of the goals for Gracemont was frequency? You really believe the 6/10W version will have the exact same frequency as Tremont?

 

[Joe NYC]

Here is a good take on that:

https://twitter.com/x/status/1447043521538527238

 

[eek2121]

Lets remember not to throw out insults and go back and forth bickering with each other. Please keep the discussion civil and avoid personal attacks, stick to the tech. Thanks. -Moderator Shmee

This. I don't agree with the mods on a lot of things (I've been banned several times, and I disagree with most of the bans , but this is one of the few times I agree with the mods), but this. Personally, in my eyes, it is okay to post your opinion (hopefully it is an educated opinion), but please, don't result to arguments and fighting when you are wrong. If you find yourself wanting to result to name calling or grandstanding in a thread for a vendor such as AMD, Apple, or Intel, just take a break. This is an Intel thread, let's talk Intel chips. On the gazillions of AMD threads (of which I, among others, are a participant of) let's talk AMD chips. On the Apple thread(s), let's talk Apple chips. If you want to talk about all three, maybe consider creating an Apple vs. Intel vs. AMD thread. I'm surprised nobody has done that yet. One of the things that has somehow, someway kept me coming back to these forums, far longer than even my user account has been registered, is the intelligent conversations that have taken place. Can we please try to keep it that way? I mean, if you dislike company X, it isn't like you even need to read the thread about company X...

I have surgery soon and will be off for a month or so...looking forward to seeing the details on all the new chip releases.

At one point we were happy that Alderlake top part might perform like a 5900X.

Why is it so bad that it might perform like a 5950X? Even if it requires 250W to do so, that's a huge jump in one generation.

When it comes to arguments, isn't that called moving goalposts?

It doesn't even look like Alderlake was aiming at desktops at all. A mobile only chip being competitive with the 5950X sounds pretty good to me. How will it do on mobile I wonder hmm?

Oh, and I won't be surprised if the Gracemont-based successor to Jasper Lake has the same 3.9GHz ST, 3.7GHz MT frequency at 10W. It won't be able to maintain 3.7GHz MT in all scenarios, probably only in certain bursty workloads. But that's the whole point of Turbo. Maximizing performance up to the limit of thermals/power.

Every generation since Goldmont it was ~30% perf/clock gain plus whatever clocks needed to get an overall gain of 50%. Goldmont Plus is an overall 50% faster than Goldmont. Tremont is an overall 50% faster than Goldmont Plus. Gracemont will probably do similar.

Have you guys also forgot one of the goals for Gracemont was frequency? You really believe the 6/10W version will have the exact same frequency as Tremont?

Gracemont will likely consume less then that. Intel has a lot of upcoming chips, and Gracemont is a big part of both ADL-S and RPL-S, if not above and beyond. I am personally genuinely looking forward to the power consumption from the small cores.





You cannot talk about the moderators and your issues with them in the public forum.
You want to talk about the specific rules your broke for your vacations?
Create an MD thread and we will tell you.


esquared
Anandtech Forum Director

 

[ondma]

And what figures are you basing all these on? Your desperation borders on grudge against your former employee to the extent that you have to pull numbers from thin air to bash the work of your betters. That's just pathetic.

Ehh... Just put him on ignore. I did it a long time ago. His hatred of Intel is so intense that any sort of reasonable discourse is impossible.

 

[Timorous]

At one point we were happy that Alderlake top part might perform like a 5900X.

I expect at 125W it will be closer to the 5900X in MT workloads.

TBH getting 5950X MT performance using a similar tier node but using 2x the power to get there is not great.

 

[Carfax83]

You can bet that if Intel does indeed catch AMD then they (AMD) will start pushing clocks and thermals if they have the leeway to do so.

Zen 3 is a proven winner, no doubt about it. Let's see how ADL stacks up.

According to Tech Spot, Intel Core i9 11980HK has significantly better power scaling than Zen 3, which I'm sure Alder Lake will also excel in. Intel has seemingly tuned their architectures for higher clock speeds more aggressively than AMD, so I doubt AMD will be able to nullify that by increasing Zen 3's clock speeds. The increased cache should help out more I'd think.

 

[BorisTheBlade82]

@Carfax83
While your argument holds true this basically just means that TigerLake is a bit less inefficient than Zen3 when both are way beyond their Perf/W sweet spot. Looking at it from this way that is not too overwhelming.

BTW
Just have a look at my PES Thread. There are samples of a 8C TGL as well as a 5900HS

 

[mikk]

Most people forget about the clock speed difference. Intel is pushing the big cores to 5.0 Ghz and the cores are getting extremely inefficient, at this speed it can't have maximum efficiency. Intel is going for the maximum performance on desktop-K, this should be more than obvious from the clock speed. 5950x on the other side is 1 Ghz down when all cores are fully loaded, but it has 16 big cores and 32 threads, it must have a much superior efficiency. This is not really a surprise, it's more of a surprise they can even reach 5950x level of performance with just 8 big cores and 8 threads less overall. ADL-S seems to be a really good first big+little attempt, from now on they can improve on this. Raptor Lake with 16 little cores is coming next year and after this we probably get updated cores in Meteor and Arrow Lake.