OS Level Throttle Fix Is Evidence Apple May Not Be Done Resolving The Issue

[RumorConsumer]

There are very few laptops on the market that can dissipate significantly more heat from the CPU, and those are large and heavy gaming + workstation machines. The MBP is able to dissipate 45-50W of CPU heat in sustained runs, which is absolutely adequate. So while these CPUs can perform better in a larger enclosure (equally true for any other CPU in the last few years), they perform within their spec.

The issues we were facing are because of power management algorithms. The throttling problem was not because the laptop can't dissipate the heat but because the power manager was misconfigured to let the CPU draw more power than the system can handle. With the firmware update, this has been fixed and the MBP now performs in line with other comparable laptops such as Razer Blade and XPS 15".

*its not a firmware update

It seems like a stop gap. And yes your point is taken

 

[Weaselboy]

Right now the fix appears to take place at the OS level and did not change the firmware version # in System Profiler.

There was a firmware update as part of that patch. Here is my post from the news thread where I mentioned it.

I installed the update on my 13" i5 model and the firmware went from 15.16.6613.0.0,0 to 15.16.6703.0.0,0, so this does install a firmware update.

I understand users in Bootcamp (Windows) are not seeing the full fix, so the fix may be a combination of a firmware and OS fix.

 

[mikethebigo]

I understand users in Bootcamp (Windows) are not seeing the full fix, so the fix may be a combination of a firmware and OS fix.

I hope the full fix doesn't require an update from Microsoft...

 

[CreeptoLoser]

Do you really think it’s fair to compare the two? ...Very different markets.

I only make the comparison to illustrate that whenever Apple needs a firmware update people go crazy for clickbait and attention. When PCs need 10 firmware updates a year nobody gives a crap because most of the public can't name hardware component makers. It's not headline grabbing.

 

[maflynn]

Dude. My PC has had eight BIOS updates from Asus this year for performance and stability issues with Coffee Lake. Those updates are issued by Intel to the OEMs so all PC motherboards are affected.

https://www.asus.com/uk/Motherboards/TUF-Z370-PLUS-GAMING/HelpDesk_BIOS/

When was the last time Apple had to release 8 firmware updates in 7 months?

Dude we are talking about Macs, which is a different animal then PCs

 

[RumorConsumer]

There was a firmware update as part of that patch. Here is my post from the news thread where I mentioned it.



I understand users in Bootcamp (Windows) are not seeing the full fix, so the fix may be a combination of a firmware and OS fix.

@Weaselboy's well stated explanation is supported by a post:
I had to reinstall macOS from internet recovery because I was missing the wallet app for Apple Pay, weird I know. Anyways, because of that I can confirm there is something more than just a firmware update going on here. Whatever Apple did to get this throttling and heat under control is definitely macOS related. The version that will download and reinstall is the original shipping version, if you run benchmarks with that it’s right back to throttle land. That’s super unfortunate. The computer should be able to cool itself, OS independent. This is getting a bit outrageous. I have a few more days before my return window is up and I’m seriously considering doing that return.

On this sister thread: https://forums.macrumors.com/thread...throttling-cpu’s-in-bootcamp-as-well.2129556/

The issue is still present in a MacOS reinstall even after its applied once. So its not sticking to hardware. Guessing that may change.

 

[mboroff]

Murmers have begun that Apple is not done working out the throttling issue.

“PCWorld’s Mah Ung speculates that even more tuning could be around the corner, and I suspect that’s because the performance uplift here, while welcome, isn’t quite where it should be yet.”

https://www.forbes.com/sites/jasone...cbook-pros-performance-problems/#2627b9359051

So here’s the thing. Right now the fix appears to take place at the OS level and did not change the firmware version # in System Profiler. This seems like an unlikely and somewhat uncharacteristicly nonpermanent way to resolve this. What seems likely/possible is that at some point soon there will be a firmware level update that resolves the issue permanently and more thoroughly under all OS conditions.

Reasons:
- this method, while helpful, is sloppy as a long term fix due to any OS needing to apply it and could possibly result in damaging the hardware which means bottom line is impacted in human service and repair hours and hardware costs. Also possible is that Apple’s strategy for issuing these types of fixes has shifted as the firmware (T2) has evolved and along with it its relationship to the OS.
- There are still some reports floating around that other methods of addressing the issue have been better performers than Apple’s fix. Yes, they might be hedging a bit on performance and heat dissipation but it’s possible they intend to hit this one more squarely than they have.
- There are people having other issues with the machine’s brain aka the T2 (kernel panics) and it’s likely this throttle issue caught Apple by surprise, the forthcoming T2 update wasn’t ready yet, this fix is now being incorporated into it and the OS level fix was the best stopgap available.
- The performance differences in benchmarks between the chips (benchmarks are a drag race, that topic has been beaten to death and still they aren’t nothing) is too small. More tuning is needed to eek out better differentiation between models. This one may be iffier than the others but I think still carries enough weight to bare mentioning.
- In every movie where the portal to hell is closed its a permanent fix not some temporary software patch.

Welcome all your thoughts either way.

So after reading all the posts on this thread, I am wondering if this fix helps running Windows 10 in a virtual machine versus using boot camp?

 

[RumorConsumer]

So after reading all the posts on this thread, I am wondering if this fix helps running Windows 10 in a virtual machine versus using boot camp?

It would be virtualized inside of an artificial hardware environment running under Mac OS X so ostensibly it would be fine.
[doublepost=1532970041][/doublepost]

So then you mean it should be fixed in macOS and not in the firmware? If it takes in account variables that can't be managed at the firmware level, then isn't the OS the proper level to fix it?
[doublepost=1532834170][/doublepost]

Is it possible that the version number of the controller we're seeing is only one specific component of the T2 and that something else was indeed updated but we haven't noticed because we do not know how to check the version of that component?
[doublepost=1532834579][/doublepost]



That's where I get skeptical and also believe that the EFI still exists but Apple simply hides it or changed the way it reports in macOS.
[doublepost=1532834953][/doublepost]Also, Craig Federighi confirmed that Mojave β will be patched. If Apple really intended to fix it at the firmware level but couldn't because the fix for the T2 was not ready, why would they bother to include in in a release that's expected in September?

That Mojave comment sticks in my craw. That they are talking about more OS level patches does not bode well for this fix being improved/transposed to "hardware" or whatever that means these days.

More on the T2 and Bridge OS assuming SMC responsibilities:

 

[RumorConsumer]

Ok after some more thought and some more research I have what I think is another, possibly more rational, explanation for believing what I wrote. The BridgeOS is so secure that it must be a royal pain in the ass to be constantly changing up unsigned Mac OS versions on it. It has so much security stuff tied to it. To me this is the clearest simplest answer that a) we havent seen an update and b) when we do it will be a good one packed full of fixes.

 

[iMacDragon]

This weeks Mojave actually updated T2 firmware, though I didn't get chance to check boot camp for any changes before my laptop was recalled for charging issue. But I suspect it wont have changed anything.

 

[RumorConsumer]

This weeks Mojave actually updated T2 firmware, though I didn't get chance to check boot camp for any changes before my laptop was recalled for charging issue. But I suspect it wont have changed anything.

Interesting.

 

[Stux]

This weeks Mojave actually updated T2 firmware, though I didn't get chance to check boot camp for any changes before my laptop was recalled for charging issue. But I suspect it wont have changed anything.

The real curious thing is if you restore HS 10.13.6 before the Supp. Update, will it still have the bad throttling behaviour now?

 

[M.Rizk]

The real curious thing is if you restore HS 10.13.6 before the Supp. Update, will it still have the bad throttling behaviour now?

Thanks to the T2 chip, I am sure no one will want to give it a shot.

 

[maflynn]

The real curious thing is if you restore HS 10.13.6 before the Supp. Update, will it still have the bad throttling behaviour now?

I'll be holding off in doing modification to the OS until 10.13.7 is out the door. I do not want to take a chance and cause problems with MBP

 

[CrashTestWalrus]

The real curious thing is if you restore HS 10.13.6 before the Supp. Update, will it still have the bad throttling behaviour now?

Yes, it does. When I did that because I had some issues with apple wallet I was told by support to use internet recovery and erase my drive and reinstall. When I did this it installed macOS prepatch and came with one ticket straight back to throttle land. My Wallet problem fixed itself, though.