Discussion Intel current and future Lakes & Rapids thread

[uzzi38]

It has, but you think Intel would play it up on their roadmaps if they were serious about launching silicon! Also if they launch it next year (see above) then that would put it in the mix with Rocket Lake and Alder Lake which would be really confusing.

I don't think any leaked roadmaps go out that far. After all, I'm at least 90% certain TGL-H is also known as TGL-R. Or at least, referenced to in drivers as TGL-R.

Whether that R means refresh or an actual -R type CPU I have no clue (wasn't the latter a BGA desktop product or something? In which case, it kinda makes sense either way...)

I agree on the timeline issue though, I can't make sense of it either. At this point I'm thinking ADL-S is a Q4 2021 product, which gives you:

CML-S - April to June - Q2 2020
RKL-S - Computex(?) - Q1 2021
ADL-S - Q4 2021
MTL-S - Q3 2022 (IMO slightly more unlikely, last three launches are mucked up by delays, I don't think 3Q cadence is normal) or Q4 2022 (return to normal yearly cadence).

 

[DrMrLordX]

I don't think any leaked roadmaps go out that far.

They've gone out to the end of 2021. Alder Lake-S definitely showed up on one of them.

 

[Exist50]

I think Intel's strategy right now is to just launch whatever's ready, whenever it's ready. They don't really have the luxury of artificially spacing out releases to give the appearance of a consistent yearly cycle. Would explain the odd cadence, at least.

 

[IntelUser2000]

UP3 RKL-U has SGX
15W RKL-U goes up to 88A (lul at potential 105W power draw at stock)

What's so funny?

You do know that Cometlake-U is already at 85A right? Icelake is at 70A. That has nothing to do with TDP. Intel chips have PL3 and PL4 values.

In mobile PL1 is king.

 

[uzzi38]

What's so funny?

You do know that Cometlake-U is already at 85A right? Icelake is at 70A. That has nothing to do with TDP. Intel chips have PL3 and PL4 values.

In mobile PL1 is king.

That's precisely what's funny. And I know it has nothing to do with TDP specifically, however what it has got to do with devices like this that are allowed to sustain their PL2s which are in excess of 70W.

That's what I find funny. Thanks to CML-U and RKL-U, we get to enjoy -H devices disguised as -U devices. It's hilarious.

Well that, and also it gives is a great taste of sustained performance under heavy load. I'll stop there before I start ranting about the gloriousness of Renoir.

 

[DrMrLordX]

I'll stop there before I start ranting about the gloriousness of Renoir.

We have other threads for that.

Devices that allow too much time above rated TDP are going to have horrible battery life and/or suffer from throttling unless the cooling solutions are excellent.

 

[IntelUser2000]

@uzzi38 What are you talking about? All 4 loads show PL1 of 45W. There's nothing in there that says its running constantly in PL2.

Exactly what DrMrLordX said.

 

[RetroZombie]

Devices that allow too much time above rated TDP are going to have horrible battery life and/or suffer from throttling unless the cooling solutions are excellent.

They could easily solve that by having different behavior when on battery or on power adapter.
I think they have done that in the past with the pentium iii mobile.

 

[DisEnchantment]

They could easily solve that by having different behavior when on battery or on power adapter.
I think they have done that in the past with the pentium iii mobile.

On my Dell XPS 13, I can easily limit this behaviour in the BIOS to make sure the 8565U does not boost to 4.6 GHz and you can get 10+ hours of battery.
When limiting this massive boost the XPS 13 can operate without the fan turned on at high rpm and the laptop is silent and quite cool. Downside is you can forget any intensive task like compiling in visual studio or run some VM.
I use it for Powerpoint and office and for remote desktop to my workstation or to ssh to my cloud VM instances

 

[lobz]

We have other threads for that.

Devices that allow too much time above rated TDP are going to have horrible battery life and/or suffer from throttling unless the cooling solutions are excellent.

BUT REAL WORLD BENCHMARKS

 

[lobz]

@uzzi38 What are you talking about? All 4 loads show PL1 of 45W. There's nothing in there that says its running constantly in PL2.

Exactly what DrMrLordX said.

I think you misunderstood him, we'll see if he elaborates.

 

[lobz]

They could easily solve that by having different behavior when on battery or on power adapter.
I think they have done that in the past with the pentium iii mobile.

I'm not sure that laptop vendors want to design cooling solutions for 'just in case you're always connected' in an ultraportable laptop. Without sufficient cooling it doesn't matter how much power you provide for your chip.

 

[Zucker2k]

I think you misunderstood him, we'll see if he elaborates.

You can't see the forest for the trees so let me help you...

Devices that allow too much time above rated TDP are going to have horrible battery life[.]

 

[lobz]

You can't see the forest for the trees so let me help you...

You must feel so good being condescending, but that's not what I meant.

 

[RetroZombie]

I'm not sure that laptop vendors want to design cooling solutions for 'just in case you're always connected' in an ultraportable laptop. Without sufficient cooling it doesn't matter how much power you provide for your chip.

You are right. But remember that intel is already 'cheating' in benchmarks with those 15W TDP u series cpus by allowing them to go up to 45W TDP. It's just extending that 'trick' when using the power adapter.

 

[Markfw]

You are right. But remember that intel is already 'cheating' in benchmarks with those 15W TDP u series cpus by allowing them to go up to 45W TDP. It's just extending that 'trick' when using the power adapter.

Not to go off-topic, but just read the new Renoir reviews.... Even at 90 watts they are losing to a 35 watt cpu.. The cheating is over ???

 

[RetroZombie]

Not to go off-topic, but just read the new Renoir reviews.... Even at 90 watts they are losing to a 35 watt cpu.. The cheating is over ???

Going to read some, haven't read one yet.
If that's the case Intel is about to start water cooling it's laptop cpus

 

[DrMrLordX]

They could easily solve that by having different behavior when on battery or on power adapter.
I think they have done that in the past with the pentium iii mobile.

There are reasons not to do that. See below.

On my Dell XPS 13, I can easily limit this behaviour in the BIOS to make sure the 8565U does not boost to 4.6 GHz and you can get 10+ hours of battery.
When limiting this massive boost the XPS 13 can operate without the fan turned on at high rpm and the laptop is silent and quite cool. Downside is you can forget any intensive task like compiling in visual studio or run some VM.
I use it for Powerpoint and office and for remote desktop to my workstation or to ssh to my cloud VM instances

At that point, though, why has the OEM provided a CPU that can boost higher? The end user wants battery life and performance, ideally. Having to choose one or the other is not so great. Also, presumably the lappie has a cooling solution that can handle the higher boost behavior, at least in bursts. Now the cooling solution is overbuilt. Some beancounter somewhere must be having fits. Poor baby.

I'm not sure that laptop vendors want to design cooling solutions for 'just in case you're always connected' in an ultraportable laptop. Without sufficient cooling it doesn't matter how much power you provide for your chip.

That's what I was sort-of thinking. OEMs are notorious for not wanting to lay out cash for extensive cooling unless it fits a majority of use cases and/or unless they can charge and arm and a leg and get away with it.

If that's the case Intel is about to start water cooling it's laptop cpus

Human bodies are full of water. The ultimate in portable computing! Sink the heat into people as they use the laptop!

 

[RetroZombie]

Human bodies are full of water. The ultimate in portable computing! Sink the heat into people as they use the laptop!

I know you are joking, but you have no idea how many people actually burned their legs because of 100ºC objects sitting on their legs for too long...

 

[RetroZombie]

Not to go off-topic, but just read the new Renoir reviews.... Even at 90 watts they are losing to a 35 watt cpu.. The cheating is over ???

Already read some, a little disappointing with the type of laptops and intel cpus they used to make the comparison.
The form factor is not the same 15"/17" vs 14", and the cpus, Linus even used one intel mobile 120 watts cpu!

They tried hard to keep intel to show up above in the charts than they ever should have, or they tried really hard to even make intel cpus more pathetic.

I don't see a reason to buy intel laptops, unless people like to pay for more than they should.

 

[Adonisds]

Are you guys ready to read many news talking about announcements and leaks of way too good to be true products because it's the first of April? I can hardly wait to witness so much creativity from the press

 

[DrMrLordX]

I know you are joking, but you have no idea how many people actually burned their legs because of 100ºC objects sitting on their legs for too long...

I was thinking of just that. But I'm sure we could rig up a system to sink water into the back as well. Lots of surface area!

 

[Zucker2k]

Rumour has it these are the final specs for Comet Lake-H. Unveils tomorrow.

Intel 10th Gen Core (Comet Lake-H) mobile series final specifications leaked - VideoCardz.com

The presentation of the upcoming Intel high-end mobile processor has leaked. Intel 10th Gen Core (Comet Lake-H) HD Tecnologia has leaked the slides for Come Lake-H processor series. The slide deck focuses on the specifications of the new series, the features over previous series (Coffee Lake)...

videocardz.com

 

[DrMrLordX]

Huh. They're going to try "thermal velocity boost" in -H CPUs?

 

[Zucker2k]

Huh. They're going to try "thermal velocity boost" in -H CPUs?

The 10980HK are undoubtedly highly binned 8 core comet lake chips so I expect thermal performance to be better than average.