Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.

Maven1975

macrumors 65816
Original poster
Aug 24, 2008
1,033
294
I have to ask because it seems to be flying under the radar.

Throttling when running W10 on my 15” 2.6 2018 MBP under bootcamp is absolutely atrocious!

Even when offloading graphics with an eGPU the CPU goes bananas.

Heck, my kids Razer Blade Stealth with 8550u (single fan 15w chip) can hold it’s temps at 75c and sustain a 3ghz boost while gaming for an extended period of time at 1440p ultra settings with the eGPU.

The MacBook goes from 4.1ghz to as low as 800mhz within a minute doing the same task.

Intel XTU stress test pegs the CPU at 100c in seconds.

Basically, the same behavior or worse that we had under OSX prior to the patch/update.
 
Heck, my kids Razer Blade Stealth with 8550u (single fan 15w chip) can hold it’s temps at 75c and sustain a 3ghz boost while gaming for an extended period of time at 1440p ultra settings with the eGPU.

It was a firmware update, so bootcamp should have been fixed.

Also, that statement is a bit nonsensical. OF course it runs cooler, it is consuming MUCH less energy... 30 less watts worth.
 
I have to ask because it seems to be flying under the radar.

Throttling when running W10 on my 15” 2.6 2018 MBP under bootcamp is absolutely atrocious!

Even when offloading graphics with an eGPU the CPU goes bananas.

Heck, my kids Razer Blade Stealth with 8550u (single fan 15w chip) can hold it’s temps at 75c and sustain a 3ghz boost while gaming for an extended period of time at 1440p ultra settings with the eGPU.

The MacBook goes from 4.1ghz to as low as 800mhz within a minute doing the same task.

Intel XTU stress test pegs the CPU at 100c in seconds.

Basically, the same behavior or worse that we had under OSX prior to the patch/update.

So you installed the update and see no difference?
 
Last edited:
It was a firmware update, so bootcamp should have been fixed.

Also, that statement is a bit nonsensical. OF course it runs cooler, it is consuming MUCH less energy... 30 less watts worth.

It is not fixed in bootcamp, it seems to rely on macos controlling it too.

Doubt I'll hear anything official, but I did raise a query about this with engineering support whilst discussing another unrelated issue.. the more people that query it the more likely they will have to do something about it.
 
It is not fixed in bootcamp, it seems to rely on macos controlling it too.

Doubt I'll hear anything official, but I did raise a query about this with engineering support whilst discussing another unrelated issue.. the more people that query it the more likely they will have to do something about it.

From MacRumors own post on the update:

"While there were many theories as to what was causing the throttling, Apple has discovered that there was a missing digital key in the firmware that impacted the thermal management system, driving down clock speeds under heavy thermal loads. This is what has been addressed in today's update."

Firmware is not OS dependent. Is there any before and after benchmarking to prove what your saying?
 
The 'firmware' in question here is the T2 with which macOS has a far deeper two way communication with than Windows, so I see nothing unfeasible about it at all.

I have myself benchmarked and watched the cpu very rapidly start doing the 800mhz bounce throttle still on windows after update but not on macos.
 
I ran a bunch of gaming tests under bootcamp yesterday and never ran any into throttling. The CPU was boosting to 4-4.5 Ghz most of the time.

P.S. Games of course won't push the CPU as much as Cinebench would...
 
  • Like
Reactions: green86
Here's powergadget readings on windows, first post patch at default settings, and then manually limiting the power usage myself much like apple is doing now on macos. VERY clear difference.
 

Attachments

  • powerdaget windows cinebench default apple post patch.PNG
    powerdaget windows cinebench default apple post patch.PNG
    32.6 KB · Views: 230
  • powerdaget windows cinebench manual 39w max limit.PNG
    powerdaget windows cinebench manual 39w max limit.PNG
    36.1 KB · Views: 256
  • Like
Reactions: g75d3 and uecker87
The 'firmware' in question here is the T2 with which macOS has a far deeper two way communication with than Windows, so I see nothing unfeasible about it at all.

I have myself benchmarked and watched the cpu very rapidly start doing the 800mhz bounce throttle still on windows after update but not on macos.

Really? Benchmarked? Before AND After with times and everything? Would you mind posting the results?

If your still seeing issues after the update playing games I'd look to the GPU.
[doublepost=1532610039][/doublepost]
Here's powergadget readings on windows, first post patch at default settings, and then manually limiting the power usage myself much like apple is doing now on macos. VERY clear difference.

Apple isn't limiting anything.
[doublepost=1532610182][/doublepost]
The 'firmware' in question here is the T2 with which macOS has a far deeper two way communication with than Windows, so I see nothing unfeasible about it at all.

Sorry, this is flat out wrong. The T2 has nothing to do with management of the CPU. Encryption, Sercure Boot, "Hey Siri", but
 
I tried Bootcamp without Intel XTU and ran into throttling issues. Once I adjusted the CPU with Intel XTU I got scores in Cinebench ranging between 1120 and around 1000 on 5 consecutive tries. Games such as Fortnite also require Intel XTU to somewhat run stable, so I think its safe to say that the Bootcamp side of this story hasn't been solved.

As a second issue is that the GPU seems to throttle hard as the chassis can't loose its heat when both the GPU and the CPU are taxed in a game. I had to set my max. Watts to 20 or lower to make sure it could keep up.
 
Lack of SMC chip in ifxit teardown, Apple differentiating SMC reset instructions specifically for T2 computers.
 
It was a firmware update, so bootcamp should have been fixed.

Also, that statement is a bit nonsensical. OF course it runs cooler, it is consuming MUCH less energy... 30 less watts worth.

The point is that the 15w Razer Stealth stays stable, has one less fan and the application being used to test does not push the CPU to the limit. (Gaming with and eGPU)
[doublepost=1532668825][/doublepost]
I tried Bootcamp without Intel XTU and ran into throttling issues. Once I adjusted the CPU with Intel XTU I got scores in Cinebench ranging between 1120 and around 1000 on 5 consecutive tries. Games such as Fortnite also require Intel XTU to somewhat run stable, so I think its safe to say that the Bootcamp side of this story hasn't been solved.

As a second issue is that the GPU seems to throttle hard as the chassis can't loose its heat when both the GPU and the CPU are taxed in a game. I had to set my max. Watts to 20 or lower to make sure it could keep up.

I have to disable the dGPU when using my internal display so those thermals are not affecting my test.

However, that’s a shame with your use case. Just makes matters worse I’m sure.

Thanks for sharing and spreading the word.
 
Would you say that the Macbook Pro is working normally in bootcamp or is it bad?
 
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.
 
  • Like
Reactions: JPack
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.

Given that Apple only spent a week to discover, test, and deploy a fix, I doubt the current fix is via firmware. It's likely just a hotfix to avoid negative coverage while they test a permanent fix on all MPB configurations.

Something so fundamental should not have slipped through preproduction testing.
 
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.


We have been having this conversation in another thread:

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

This thread is actually a great compliment and this post answers the question we were ending up with. @johnalan you appear to be off the hook. @CrashTestWalrus would also be good to know the exact reported version in System Information.

From what I am gathering, the T2 is now the replacement for all that the SMC used to cover (power management, fan control, etc) and is a security guard for the EFI firmware. What we're observing is the interplay between the OS and the T2 but the T2 itself doesn't appear to have received an update. Perhaps thats what we're waiting for.
[doublepost=1532957923][/doublepost]
Totally agree. This is my current trust test. I equally ignore people who believe it to be DOOM and people who believe it be excusable. It’s neither, but whatever allowed it to happen must be corrected. It’s not acceptable.
Yeah hard to straddle the line. I both agree its a big lapse and have faith they will rectify to an appropriate level (more than they are now).

Its definitely on somebody's performance review.
[doublepost=1532957961][/doublepost]
Many thanks for the honest answer.

I’m going to wait for the 2019 version now .....
A little dramatic I think. Just wait for a hardware level resolution and get your machine.
[doublepost=1532958424][/doublepost]
Also everything SMC used to deal with.
Do you have a source for that info?

https://www.macworld.com/article/32...e-imac-pro-the-start-of-a-mac-revolution.html

Because that doesn't appear to be the case.


I found some more references to the T2 being the full replacement to the SMC. It seems to be universally reported in Snell's writing and Cabel Sasser, Panic cofounder:

Screen Shot 2018-07-30 at 12.45.34 PM.png



So a friend asked me what I thought about the issue returning after the shipping build of Mac OS was reinstalled and if this was a bad sign. I don't think its bad. I hypothesize that its Apple leveraging the properties of this new Bridge OS and how flexible it is to apply a short term fix like this from the OS side. Isn't what they are officially saying is that there are keys missing on the Mac OS side from the way these two pieces interface? I would suggest, again, that this is not the permanent fix and that Apple would want to resolve this "in hardware" or as close to "in hardware" as we get these days ie sufficiently so that Bridge OS retains the fix when the supplemented Mac OS version isnt present.

Bridge OS is an entirely new kind of software to be managing those aspects of operation but its higher order than anything the SMC ever operated on. At this point, it would seem a safe bet that the functions previously held by the SMC were effectively "ported" to Bridge OS and this is a glitch in the port.

Pretty clear. Its more or less what he said in that article.
 
Last edited:
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.