thermal throttling More details about the throttling issues of the 15" core i9 MacBook Pro, this time with Final Cut Pro X

[Crunchy Dragon]

8 minutes ago, 8uhbbhu8 said:

They didn't even bother to engineer a proper cooling solution for a chip THEY KNOW is a heat magnet

It's not even just that, MacBooks have pretty terrible cooling in general.

[8uhbbhu8]

2 minutes ago, Crunchy Dragon said:

It's not even just that, MacBooks have pretty terrible cooling in general.

That's very true. At least with most of the others they at least tried somewhat to keep it cool. And they chose cpus that were decent for that formfactor at the time.

[Crunchy Dragon]

Just now, 8uhbbhu8 said:

That's very true. At least with most of the others they at least tried somewhat to keep it cool. And they chose cpus that were decent for that formfactor at the time.

Making them thinner and more compact doesn't do much to help them either, everybody knows Coffee Lake runs hot.

[8uhbbhu8]

1 minute ago, Crunchy Dragon said:

Making them thinner and more compact doesn't do much to help them either, everybody knows Coffee Lake runs hot.

Yeah Coffee Lake Never was a great idea to put in it in the first place with that cooling capacity. Doesn't matter what you do with pipes and fans. Metal can only dissipate a certain amount of heat before it's overwhelmed. You'd think they would have been able to calculate that and realize it was a bad idea. Although this IS Apple so take from that what you will.

[captain_to_fire]

35 minutes ago, Crunchy Dragon said:

Making them thinner and more compact doesn't do much to help them either, everybody knows Coffee Lake runs hot.

That's why I think Coffee Lake is more of a half-assed response to Ryzen imo

[Crunchy Dragon]

Just now, captain_to_fire said:

That's why I think Coffee Lake is more of a half-assed response to Ryzen imo

I wouldn't go that far, the main problem is price and heat output. It's not entirely half-baked.

[leadeater]

4 minutes ago, Crunchy Dragon said:

It's not entirely half-baked.

Well it is fully baked that's for sure 

[captain_to_fire]

Just now, leadeater said:

Well it is fully baked that's for sure 

I'll wait until an article appears online saying that Intel's core i9 contributes greatly to global warming and environmental activists with protest on Intel HQ saying "i9 kills polar bears"

7 minutes ago, Crunchy Dragon said:

I wouldn't go that far, the main problem is price and heat output. It's not entirely half-baked.

If that's the case then Intel's hope is to use 10nm transistors which at the moment they're having troubles. I think Cannonlake with 10nm is only for mobile devices.

[D13H4RD]

Coffee Lake was essentially using the best of what they currently had to make it as competitive as possible to Ryzen on an overall performance level. 

 

We won't see meaningful improvements to heat and power efficiency until Intel rolls out a drastically overhauled core architecture 

[mr moose]

3 hours ago, captain_to_fire said:

How I wish there's an Android phone that receives security patches and new Android versions for more than 36 months and doesn't require rooting. Tbh I prefer Samsung's hardware design than Apple or Google but those two offer better software experience imo.

If I could keep a phone alive for more than 12 months I'd know if that was true or not, but my last to androids have received security updates almost monthly the whole time I had them.

 

TBH I'm not really sure why people keep posting that picture for security update claiming only 6% get them or something, that's never been my experience and I refuse to accept I am always in that 6%.

[Guest]

1 minute ago, mr moose said:

If I could keep a phone alive for more than 12 months I'd know if that was true or not, but my last to androids have received security updates almost monthly the whole time I had them.

 

TBH I'm not really sure why people keep posting that picture for security update claiming only 6% get them or something, that's never been my experience and I refuse to accept I am always in that 6%.

Most people buy budget phones, you get what you pay for. 

 

1 hour ago, leadeater said:

Well it is fully baked that's for sure 

It's not the only thing thats fully baked  

[mr moose]

Just now, RorzNZ said:

Most people buy budget phones, you get what you pay for. 

 

It's not the only thing thats fully baked  

I do buy budget phones,   The most expensive phone I bought was a windows phone which was $199.  the rest have all been around the $90-$160 mark.

[captain_to_fire]

8 minutes ago, mr moose said:

TBH I'm not really sure why people keep posting that picture for security update claiming only 6% get them or something, that's never been my experience and I refuse to accept I am always in that 6%.

I think those pie charts/bar graphs refer to Android versions and not security patches. Google decided to separate security patches from Android versions. Apple on the other hand combines new iOS version with security patches that's why the latest stable build is iOS 11.4.1 whereas with Google it's Android Security Patch Level 2018-07-05. I think monthly patches are expected from high/mid tier Android phones depending on the vendor. I doubt those $50-100 Android phones from China will ever see even a single update.

[Guest]

Just now, mr moose said:

I do buy budget phones,   The most expensive phone I bought was a windows phone which was $199.  the rest have all been around the $90-$160 mark.

Then I guess you are most people and 6% is inaccurate

[DKL]

As someone that grew up in a family of Mac users, I'm not the least surprised by this. I owned a Macbook that overheated like crazy all the time. My mother's old one is still sitting around in her house. An unused brick since it has always overheated like crazy. Macbook Pros were never immune to this 'feature' either. Feeling like my computer was burning my lap was always a very fun experience though..

Macbook: The computer that requires a laptop cooling pad just to make use of some performance.. and even then still shouldn't be left on for too long. 

[DanielMDA]

On 19/7/2018 at 5:57 AM, ImpulseRez said:

A bit ridiculous to blame Intel on this one. Apple must have made a conscious design choice to keep the laptop thin at the expense of cooling. Intel chips are pretty good from an efficiency perspective (better than AMD at least), and being the first 6 core laptop chips are always going to be thermally dense to start with.

Also, bringing up the ARM mention is pretty absurd (performance compared to an 8th Gen 6 core Intel CPU is not going to be in the same league!)

Yeah, Intel chips are great at efficiency, but I disagree on the second part, AMD Ryzen CPUs are just as good if not more efficient than Intel´ps now. Specially with APUs and now 12nm. You can manage to put an AMD Ryzen 7 2700/1700 on a Laptop and not overheat, and the Raven Rifge APUs maintiain their 15W TDP even with the Vega GPU at work.

[Stefan Payne]

3 hours ago, Crunchy Dragon said:

It's not even just that, MacBooks have pretty terrible cooling in general.

While that might be true, it looks like this CPU is over 45W TDP and that the TDP Rating isn't really correct...

 

I've looked up some Intel Processors and somethings just don't add up:

Just scroll down and look at that.

https://www.tomshw.de/2018/04/19/ryzen-2000-performace-leistungsaufnahme-temperaturen/17/

 

No need for translation and the reason why nobody mentions power consumption since Ryzen in recommendation pages. Because the Efficiency of the Intel CPUs isn't that great...

 

Another thing:

https://www.anandtech.com/show/12841/amd-preps-new-ryzen-2000series-cpus-45w-ryzen-7-2700e-ryzen-5-2600e

Although that seems like DESKTOP Processors and not mobiles.

 

Money Quote:

2.8GHz is the specified Base frequency of the 8 Core/16 Thread CPU, 3.1GHz of the 6 Core/12 Thread CPU...

And then they specifiy the "45W" with 2.9GHz and up to 4.6GHz Boost...

 

----------------

now back to Intel:

 

 

What is interesting, that the 35W TDP Desktop processor clock far lower. Like 2.3GHz Base tops and 3.6GHz Boost.

Like the i7-8600T:

https://geizhals.de/intel-core-i5-8600t-cm8068403358708-a1796008.html?hloc=de

Another interesting part of Information is the 25W TDP Downclock that is mentioned here.

 

And the "45W" TDP mobile looks astonishingly like the 65W TDP Desktop i7-8700 (non K Version. K is officially 95W TDP).

https://geizhals.de/intel-core-i7-8700-bo80684i78700-a1797792.html?hloc=de

 

And we also have 45W TDP Desktop processors, though they are Xeon:

https://geizhals.de/intel-xeon-e3-1260l-v5-cm8066201921903-a1330448.html?hloc=de

2.9GHz Base, 3,9GHz Boost, but only Skylake -> 4C/8T...

 

 

That does not really add up on Intels side and looks a bit optimistic

 

The only thing we are able to know more is to get one of those mobile CPUs, modify the Board and measure the power consumption...

But looking at the Desktop versions, we can say that the mobile ones look weird...

 

Especially if you compare the mobile with the 45W Skylake Xeon. 

The 45W Mobile Part has higher boost, same base and same TDP as a 4Core/8 THread CPU, although the architecture is basically the same and the manufacturing still on the same node...

[Stefan Payne]

29 minutes ago, DanielMDA said:

Yeah, Intel chips are great at efficiency, but I disagree on the second part, AMD Ryzen CPUs are just as good if not more efficient than Intel´ps now. Specially with APUs and now 12nm. You can manage to put an AMD Ryzen 7 2700/1700 on a Laptop and not overheat, and the Raven Rifge APUs maintiain their 15W TDP even with the Vega GPU at work.

AMD CPUs right now are more efficient and consume less power than the Intel.

That is why nobody mentions the better Efficiency/lower Power Consumption of the CPU in recommendation Threads these days

On 19.7.2018 at 12:57 PM, ImpulseRez said:

A bit ridiculous to blame Intel on this one. Apple must have made a conscious design choice to keep the laptop thin at the expense of cooling. Intel chips are pretty good from an efficiency perspective

Everything you stated here is false. And provable so.

From what we know right now, this chip is a bit over 45W TDP, and it looks like its a pretty optimistic TDP Rating.

 

On 19.7.2018 at 12:57 PM, ImpulseRez said:

(better than AMD at least),

That is just not true.

Look here:

https://www.tomshw.de/2018/04/19/ryzen-2000-performace-leistungsaufnahme-temperaturen/17/

 

Power Consumption of Ryzen 2700x is equal to i3-8350k

 

On 19.7.2018 at 12:57 PM, ImpulseRez said:

and being the first 6 core laptop chips are always going to be thermally dense to start with.

No, not if you do it right, and spec the CPU according to its regular power consumption or reduce Clockrate and voltage until it fits.

What you miss is the i7-8600T wich is a 35W TDP CPU. 

 

On 19.7.2018 at 12:57 PM, ImpulseRez said:

Also, bringing up the ARM mention is pretty absurd (performance compared to an 8th Gen 6 core Intel CPU is not going to be in the same league!)

you don't have any proof of that as the cross plattform benchmarks are pretty much impossible to do.

And the ARM Chips are developed for a total different usage pattern anyway, so it doesn't make too much sense to do that anyway...

[D13H4RD]

29 minutes ago, Stefan Payne said:

From what we know right now, this chip is a bit over 45W TDP, and it looks like its a pretty optimistic TDP Rating.

It’s not “a little”.

 

Depending on the clockspeed and probable voltage, it can either be that or a lot.

 

For all we know, 45W is the probable TDP at base 

[mr moose]

1 hour ago, Stefan Payne said:

While that might be true, it looks like this CPU is over 45W TDP and that the TDP Rating isn't really correct...

 

I've looked up some Intel Processors and somethings just don't add up:

 

Especially if you compare the mobile with the 45W Skylake Xeon. 

The 45W Mobile Part has higher boost, same base and same TDP as a 4Core/8 THread CPU, although the architecture is basically the same and the manufacturing still on the same node...

Stop talking about TDP,  you have been told by many people what it is and why it is an accurate spec,  but you insist on telling people it is wrong. 

1 hour ago, D13H4RD2L1V3 said:

It’s not “a little”.

 

Depending on the clockspeed and probable voltage, it can either be that or a lot.

 

For all we know, 45W is the probable TDP at base 

It is the TDP because TDP is measured at base clock.  Forget everything that Stefan says, because this has nothing to do with efficiency or xeons or cores, but purely about the lack of cooling in the mac. 

[D13H4RD]

4 minutes ago, mr moose said:

It is the TDP because TDP is measured at base clock.  Forget everything that Stefan says, because this has nothing to do with efficiency or xeons or cores, but purely about the lack of cooling in the mac. 

And other compact laptops with inadequate cooling.

 

I could talk about the 8950HK’s heat output and Coffee Lake mobile’s limits in pushing performance and achieving a balance between it and efficiency, but that’s for another time (and Intel would probably be rolling out a more efficient architecture by that time.)

 

The thing is that with Coffee Lake, it follows that TDP definition exactly. You can’t reasonably expect the CPU to maintain the same characteristics as Kaby Lake with 2 extra cores and more aggressive clocks, especially when the core architecture has not changed too much. We can talk about how Intel could’ve made it more efficient or something but Coffee Lake is already out. The least manufacturers can do right now is implement cooling systems which can handle the increased heat demands or just not offer them if it’s unsuitable 

[Stefan Payne]

7 minutes ago, mr moose said:

Stop talking about TDP,  you have been told by many people what it is and why it is an accurate spec,  but you insist on telling people it is wrong. 

It is the TDP because TDP is measured at base clock.  

Its not my Problem that Intel uses such a shitty definition of TDP.

And why are you also defending Intel right now?

 

So you are just making things up like you did in the other thread with the "TV" thing, where you made up your own definition. And now you are doing the same with the TDP.

 

If Intel is not able to tell the Notebook Manufacturers what TDP to expect and cool, it is their fault. And that is what it looks like because all CPUs get really realyl hot.

 

 

Forget everything that Stefan says, because this has nothing to do with efficiency or xeons or cores, but purely about the lack of cooling in the mac. 

Why are you defending Intel so much and try to deflect the blame from them, when it is clear that they are not innocent in this mess?!

[mr moose]

15 minutes ago, Stefan Payne said:

Its not my Problem that Intel uses such a shitty definition of TDP.

The problem is not the definition, it's that you don't understand it.

Quote

And why are you also defending Intel right now?

Call it defense, but if you weren't lying about it all the time no one would say anything.

 

Quote

So you are just making things up like you did in the other thread with the "TV" thing, where you made up your own definition. And now you are doing the same with the TDP.

 

Please, Made up? you think  @D13H4RD2L1V3 @leadeater myself and a few others are just making this up?

 

 

Quote

If Intel is not able to tell the Notebook Manufacturers what TDP to expect and cool, it is their fault. And that is what it looks like because all CPUs get really realyl hot.

Intel are able to tell them what the TDP.

 

Quote

Why are you defending Intel so much and try to deflect the blame from them, when it is clear that they are not innocent in this mess?!

Why are you lying about it so much and trying to blame them for it?

[leadeater]

2 minutes ago, mr moose said:

Please, Made up? you think  @D13H4RD2L1V3 @leadeater myselkf and a few others are just making this up?

Shh you'll give the game away.

 

Spoiler

*Secret handshake for those in the know*

 

[D13H4RD]

Also, the watts used in TDP are a measure of heat.

 

Watts in TDP does not necessarily equal to Watts in package power