The No.1 Website for Pro Audio
 Search This Thread  Search This Forum  Search Reviews  Search Gear Database  Search Gear for sale  Search Gearslutz Go Advanced
MBP High Sierra and Mojave CoreAudio overload bug Audio Interfaces
Old 28th August 2018
  #1
Gear Head
MBP High Sierra and Mojave CoreAudio overload bug

I’m looking for other MacBook Pro users to verify a CoreAudio overload bug that occurs in both High Sierra 10.13.6 and the latest Mojave Beta 10.14. I’ve tested a new MBP 2018 with the i9 cpu, 32GB ram, 1TB HDD, and a late 2013 MBP & 2011 MBP. All running High Sierra.

The issue involves a background kernel process called the AppleSmartBatteryManager. On my systems, it runs a diagnostic poll of the battery status every 60 seconds. Regardless of being plugged into the AC adapter. This event creates a CPU spike that can overload your CoreAudio and cause a dropout/glitch sound. Typically because most pro audio applications are putting your active track on the same core as the background kernel processes.

I’ve created a couple videos detailing my steps, this being the latest with Mojave, the other video link is in its YouTube description. There I also reproduce it within High Sierra: YouTube

Here are the steps if you don’t want to watch the video. I use Logic Pro to demonstrate it since its CPU meter isn’t averaged like in Ableton 10 or Cubase 9.5.3. It’s easier to see the cpu spike. I have tested Ableton and Cubase and can reproduce the overload events.

1. Open Logic
2. Create Instrument Track with Alchemy Bowed Metal Space patch (or some other cpu intensive plugin) I usually add in a default chromaverb fx to get the cpu on live play over 50%. I’ve also used Keyscape, Kontakt libraries, and uh-e synths with similar results.
3. Double click your Logic CPU meter to get floating window showing all cores.
3. Open your OSX Console App and search Battery and note the time the AppleSmartBatteryManager start poll event occurs.
4. Open your System Pref Time and Date leave that viewable to track time.
5. Start playing the Alchemy patch live as the active track and watch CPU meter as time approaches the AppleSmartBatteryManager poll start.
6. You should encounter a sizable cpu spike, and high potential for a audio drop out. Repeat a few more min cycles noting if you got a glitch audio dropout.
7. Return to the System Console and search those event times. You should see the AppleSmartBatteryManager poll start, then immediately restart, then some of its command checks, and if you had a dropout, a CoreAudio overload message(s). All occurring at the same second mark.

On my tested systems I get a CoreAudio overload event more than 90% of the time. And register a CPU spike every time. I’ve tested with the internal MBP sound, and RME Babyface Pro, and a Digigrid D Ethernet Audio Interface. Depending on project size this can occur at all CoreAudio buffer settings. My example above I had 128 selected. Changing the processing core setting in Logic has no effect. It always puts the active track the same as the kernel level processes.

I’ve started a bug report through Apple developer for Mojave. But as with all things, the more eyes that get on this the more likely a fix.
Old 28th August 2018
  #2
Mojave is no where near ready for DAW and audio interfaces. My Motu Ultralite Mk3s dont even have recognized inputs. Outs are fine but ins are not. I will wait until public release but even then I am going to have a boot drive with high sierra standing by.
Old 28th August 2018
  #3
Gear Head
Quote:
Originally Posted by Blacklilyorchard View Post
Mojave is no where near ready for DAW and audio interfaces. My Motu Ultralite Mk3s dont even have recognized inputs. Outs are fine but ins are not. I will wait until public release but even then I am going to have a boot drive with high sierra standing by.
Are you running a MBP? My bug is also on High Sierra. The Mojave test was just to see if they fixed it. I’d be curious to know if you see the cpu spike and can reproduce the CoreAudio dropout.
Old 29th August 2018
  #4
Gear Head
I've gotten to the root of the issue and submitted an Apple Dev Bug Report. The AppleSmartBatteryManager driver is part of the IOKit framework, as is CoreAudio. The IOKit has an associated WorkLoop which is basically a que of background tasks it runs. When these are single threaded, like the polling method the Battery Manager uses for its CommandGate checks, or running the active live track in Logic, they exist on the same core, and basically get lined up to run in the WorkLoop. However, the Battery Manager gets priority based on its Command checks. It locks out the core for its cycle. So when a low latency single threaded coreaudio event happens at the same time, you get the skipped cycle error, and then the coreaudio stream can overload producing the audio dropout. You can avoid it by running a higher CoreAudio buffer sample, but if the CPU load is high enough on your audio process, it can still conflict and overload.

Apple did respond for additional diagnostic info, and I also spoke with Logic support and gave them the same details. I even spoke with RME since I initially thought it was exclusive to their driver. Hopefully it gets addressed in some form. To me moving it off the shared WorkLoop, or giving CoreAudio priority should solve it. I verified the bug by disabling the battery manager and the problem went away. Not a fix though, since you cannot get battery health status and have safe guards against overheating or overcharging.
Old 29th August 2018
  #5
Gear Maniac
Nice job thanks
Old 6th September 2018
  #6
Gear Nut
 

@zenwolf: really valuable info, thank you very much !
Old 6th September 2018
  #7
Quote:
Originally Posted by zenwolf View Post
Are you running a MBP? My bug is also on High Sierra. The Mojave test was just to see if they fixed it. I’d be curious to know if you see the cpu spike and can reproduce the CoreAudio dropout.

quad i7 5k iMac late 2015. so no battery or battery manager

Last edited by Blacklilyorchard; 6th September 2018 at 09:23 PM..
Old 26th September 2018
  #8
Here for the gear
Quote:
Originally Posted by zenwolf View Post
I’m looking for other MacBook Pro users to verify a CoreAudio overload bug that occurs in both High Sierra 10.13.6 and the latest Mojave Beta 10.14. I’ve tested a new MBP 2018 with the i9 cpu, 32GB ram, 1TB HDD, and a late 2013 MBP & 2011 MBP. All running High Sierra.

The issue involves a background kernel process called the AppleSmartBatteryManager. On my systems, it runs a diagnostic poll of the battery status every 60 seconds. Regardless of being plugged into the AC adapter. This event creates a CPU spike that can overload your CoreAudio and cause a dropout/glitch sound. Typically because most pro audio applications are putting your active track on the same core as the background kernel processes.

I’ve created a couple videos detailing my steps, this being the latest with Mojave, the other video link is in its YouTube description. There I also reproduce it within High Sierra: YouTube

Here are the steps if you don’t want to watch the video. I use Logic Pro to demonstrate it since its CPU meter isn’t averaged like in Ableton 10 or Cubase 9.5.3. It’s easier to see the cpu spike. I have tested Ableton and Cubase and can reproduce the overload events.

1. Open Logic
2. Create Instrument Track with Alchemy Bowed Metal Space patch (or some other cpu intensive plugin) I usually add in a default chromaverb fx to get the cpu on live play over 50%. I’ve also used Keyscape, Kontakt libraries, and uh-e synths with similar results.
3. Double click your Logic CPU meter to get floating window showing all cores.
3. Open your OSX Console App and search Battery and note the time the AppleSmartBatteryManager start poll event occurs.
4. Open your System Pref Time and Date leave that viewable to track time.
5. Start playing the Alchemy patch live as the active track and watch CPU meter as time approaches the AppleSmartBatteryManager poll start.
6. You should encounter a sizable cpu spike, and high potential for a audio drop out. Repeat a few more min cycles noting if you got a glitch audio dropout.
7. Return to the System Console and search those event times. You should see the AppleSmartBatteryManager poll start, then immediately restart, then some of its command checks, and if you had a dropout, a CoreAudio overload message(s). All occurring at the same second mark.

On my tested systems I get a CoreAudio overload event more than 90% of the time. And register a CPU spike every time. I’ve tested with the internal MBP sound, and RME Babyface Pro, and a Digigrid D Ethernet Audio Interface. Depending on project size this can occur at all CoreAudio buffer settings. My example above I had 128 selected. Changing the processing core setting in Logic has no effect. It always puts the active track the same as the kernel level processes.

I’ve started a bug report through Apple developer for Mojave. But as with all things, the more eyes that get on this the more likely a fix.
Hi, I'm another 2018 mbp 15" user here (core i7 2.6, 16ram, 560x, 1tbssd)
How is it going with your setup? I've noticed that it has got much better than before after supplemental update, currently at release version of mojave, though, issue, even if much much minor (by really far) is still present. Probably it isn't the same one as before.
p.s. I've just got zoom tac 2r thunderbolt audio interface and an tb3 to tb2 adapter to see how it goes, eventhough it happens really rarely (once every few minutes) during just normal iTunes playback, still it's there...
Very slight and fast (not as before) but still, annoying, especially considering that 2016's model wasn't affected by that problem. How is yours instead? And how did you connected ethernet cable? (i mean which adaptors)
Old 26th September 2018
  #9
Gear Head
Quote:
Originally Posted by 8SlaiN8 View Post
Hi, I'm another 2018 mbp 15" user here (core i7 2.6, 16ram, 560x, 1tbssd)
How is it going with your setup? I've noticed that it has got much better than before after supplemental update, currently at release version of mojave, though, issue, even if much much minor (by really far) is still present. Probably it isn't the same one as before.
p.s. I've just got zoom tac 2r thunderbolt audio interface and an tb3 to tb2 adapter to see how it goes, eventhough it happens really rarely (once every few minutes) during just normal iTunes playback, still it's there...
Very slight and fast (not as before) but still, annoying, especially considering that 2016's model wasn't affected by that problem. How is yours instead? And how did you connected ethernet cable? (i mean which adaptors)
The problem is still there and it’s really an OS X issue rather than hardware. Though MacBooks have the specific problem of the smart battery driver, anyone on a Mac can experience poor audio performance. #1 thing to do is create a separate user for your pro audio work. On that user do not setup iCloud. Turn off Siri. Disable Notifications. Make sure you don’t have Bluetooth PAN as a Network Device. Then get the free app called Onyx and use it to turn off Spotlight indexing. Also stay offline while working in your DAW. That’ll get you mostly there. Other things to do are a bit more involved and require disabling system files. You can prevent the battery driver from loading and disable other background processes. It’s a pain, but this issue is going to require some major code rewrite from Apple. The system is powerful for sure, it just isn’t designed around low latency audio. Real-time low latency audio processes are single threaded onto the same core that the system bus background tasks run and often those are either a higher priority than your coreaudio stream or they run a special type of process that essentially locks the cpu core until all of its commands have completed. Those are the usual culprits for overloading the audio buffer. The battery manager being the worst.
Old 26th September 2018
  #10
Here for the gear
Quote:
Originally Posted by zenwolf View Post
The problem is still there and it’s really an OS X issue rather than hardware. Though MacBooks have the specific problem of the smart battery driver, anyone on a Mac can experience poor audio performance. #1 thing to do is create a separate user for your pro audio work. On that user do not setup iCloud. Turn off Siri. Disable Notifications. Make sure you don’t have Bluetooth PAN as a Network Device. Then get the free app called Onyx and use it to turn off Spotlight indexing. Also stay offline while working in your DAW. That’ll get you mostly there. Other things to do are a bit more involved and require disabling system files. You can prevent the battery driver from loading and disable other background processes. It’s a pain, but this issue is going to require some major code rewrite from Apple. The system is powerful for sure, it just isn’t designed around low latency audio. Real-time low latency audio processes are single threaded onto the same core that the system bus background tasks run and often those are either a higher priority than your coreaudio stream or they run a special type of process that essentially locks the cpu core until all of its commands have completed. Those are the usual culprits for overloading the audio buffer. The battery manager being the worst.
Thanks for these tips! I'll do, though it's a big pita because before on Sierra it was flawless... at least regards streaming audio...
Anyway, sounds like it isn't issue with T2 chip not being powerful enough but with the whole system, so it's more likely that it's gonna be fixed soon. (hopefully with enough care about all of audio guys on osx)
Did these steps you've provided give you good enough performance?
Old 26th September 2018
  #11
Here for the gear
 

Quote:
Originally Posted by zenwolf View Post
I’ve started a bug report through Apple developer for Mojave.
Me too, did they ask you to do a "tailspin"? I can trigger it but no corresponding file appears?
Old 27th September 2018
  #12
Gear Head
Quote:
Originally Posted by amok_san View Post
Me too, did they ask you to do a "tailspin"? I can trigger it but no corresponding file appears?
They had me run a Sysdiagnose Report I also gave specifics on the occurrence with my MBP. No telling if they will actually address the bug, but they did respond twice -- although that could have been automated.

Yes, my fixes greatly improved performance. Logic however is the most finicky of the DAWs for background processes causing issues. Ableton is pretty solid with the fixes. I have a RME BFP and a Digigrid D, and the Digigrid tends to be more stable, even working fine without the fixes, but causes the system fan to run higher. I think that is a firmware issue on the MBP relating network traffic to fan speeds. It does the same when iCloud processes are running. Perhaps it is the network audio vs usb protocol.

I use a paid program called TinkerTool System to change a setting that gives priority to background network traffic, so that may be helping.

Every once in a while I'll still get a glitch on the RME, but it is very rare. Some system process is still overloading the buffer, but that is only at a low buffer setting of 64 samples. 128 and higher don't have issues. About to try out a Apollo X6 Thunderbolt 3 to see if it does better, I'll post my results. I should have that interface next week.
Old 28th September 2018
  #13
Here for the gear
Quote:
Originally Posted by zenwolf View Post
They had me run a Sysdiagnose Report I also gave specifics on the occurrence with my MBP. No telling if they will actually address the bug, but they did respond twice -- although that could have been automated.

Yes, my fixes greatly improved performance. Logic however is the most finicky of the DAWs for background processes causing issues. Ableton is pretty solid with the fixes. I have a RME BFP and a Digigrid D, and the Digigrid tends to be more stable, even working fine without the fixes, but causes the system fan to run higher. I think that is a firmware issue on the MBP relating network traffic to fan speeds. It does the same when iCloud processes are running. Perhaps it is the network audio vs usb protocol.

I use a paid program called TinkerTool System to change a setting that gives priority to background network traffic, so that may be helping.

Every once in a while I'll still get a glitch on the RME, but it is very rare. Some system process is still overloading the buffer, but that is only at a low buffer setting of 64 samples. 128 and higher don't have issues. About to try out a Apollo X6 Thunderbolt 3 to see if it does better, I'll post my results. I should have that interface next week.
I've tried it with zoom tac-2r (just bought it to test thunderbolt performance) and apart from gaining some more speed in roundtrip performance is almost the same, can't really tell the difference, very rare stutters... also, looking forward maybe newest logic 10.4.2 will fix smth... anyway, looking forward to to your report about uad tb3 interface
Old 29th September 2018
  #14
Lives for gear
 
stratology's Avatar
 

Quote:
Originally Posted by zenwolf View Post
..thing to do is create a separate user for your pro audio work. On that user do not setup iCloud. Turn off Siri. Disable Notifications. Make sure you don’t have Bluetooth PAN as a Network Device. Then get the free app called Onyx and use it to turn off Spotlight indexing. It also indicates a fundamental misunderstanding of how Spotlight actually works - there is no scheduled indexing that could actually affect performance.
Also stay offline while working in your DAW. That’ll get you mostly there. Other things to do are a bit more involved and require disabling system files. You can prevent the battery driver from loading and disable other background processes.

No. What you describe ranges from completely unnecessary to actually damaging to a DAW machine. When you disable Spotlight indexing, Logic will not work correctly. It also indicates a fundamental lack of understanding of how Spotlight actually works - there is no scheduled indexing that could have a negative effect on performance.
Turning off networking affects updates and the help system. Apps like Onyx are unnecessary and usually do more harm than good.


Messing with system files and background processes is a truly bad idea.
If you have messed with system files on your beta testing system, it also invalidates any Radar that you may have submitted when beta testing Mojave.

Last edited by stratology; 29th September 2018 at 02:41 AM..
Old 29th September 2018
  #15
Lives for gear
 
maxwelldub's Avatar
 

I'm experiencing this glitch as well. I had a 2015 maxed out 15" MacBook Pro + 2018 13" maxed out MacBook Pro (except 256gb storage) and currently have a base model 2018 15" MacBook Pro that Apple swapped for me. This first showed up months ago after applying the spectre/meltdown patch on my 2015 MacBook Pro so I assume its related to that. Any clue when this will be fixed? I'm thinking of calling AppleCare and referencing the case number...
Old 29th September 2018
  #16
Gear Head
Quote:
Originally Posted by stratology View Post
That's a little misleading. You're a beta tester, possibly found a bug, and reported it. This has nothing whatsoever to do with using a current OS release and hardware with drivers that are actually qualified for this OS release. No. What you describe ranges from completely unnecessary to actually damaging to a DAW machine. When you disable Spotlight indexing, Logic will not work correctly. It also indicates a fundamental lack of understanding of how Spotlight actually works - there is no scheduled indexing that could have a negative effect on performance.
Turning off networking affects updates and the help system. Apps like Onyx are unnecessary and usually do more harm than good.


Messing with system files and background processes is a truly bad idea.
If you have messed with system files on your beta testing system, it also invalidates any Radar that you have submitted.
I specified it was at your own risk. And it shows how little you know of the iokit framework and overall OS architecture. It is quite simply a system bus event taking priority over a real-time single threaded coreaudio event, and/or said event(s) performing a commandgate lock preventing asynchronous event scheduling. CoreAudio and SMBus events share the same recursive workloop and get threaded on the same core. My bug report was based on a clean install of both High Sierra and Mojave Beta with no alterations. I simply said turn off WiFi while working in a DAW, and disable unused networking devices like Bluetooth PAN. All things I mentioned can easily be reactivated. If you’re not comfortable in terminal and disabling things then by all means wait for an Apple official fix, although this issue has been around for years.
Old 29th September 2018
  #17
Gear Head
FWIW this began with RME support who provided a diagnostic driver to see if they might be causing the glitch many users were reporting. We were testing current High Sierra. Through reading the console events while using their driver it became obvious the same battery event was triggering the coreaudio buffer overload using small buffer sizes of 128 or 64 samples. We tested through RME’s guidance running their devices in Class Compliant mode and confirmed the issue was OS X related. I merely tested the beta Mojave to see if the issue was still present. Not for driver compatibility.
Old 29th September 2018
  #18
Lives for gear
 
stratology's Avatar
 

Quote:
Originally Posted by zenwolf View Post
My bug report was based on a clean install of both High Sierra and Mojave Beta with no alterations.
OK, that's good practice.

Quote:
Originally Posted by zenwolf View Post
I simply said turn off WiFi while working in a DAW, and disable unused networking devices like Bluetooth PAN. All things I mentioned can easily be reactivated. If you’re not comfortable in terminal and disabling things then by all means wait for an Apple official fix, although this issue has been around for years.
Two different issues here - your bug report and good relating information, and bad general advice about system maintenance.


It's not about being 'comfortable', it's about messing up the functionality of your system. Again, when Spotlight indexing is disabled, LPX will not work correctly because it uses Spotlight. If a network device is not in use, it does not affect performance in any way. Things like 'System cleaning' in Onyx are unspecific, and completely unnecessary. If you do things like log 'cleaning' in Onyx, your sysdiagnose report will not have the relevant information that you deleted. Logs are useful, and in the rare case that something goes wrong and a runaway error creates large log files, the better option is to delete the affected log manually.
Old 29th September 2018
  #19
Gear Head
Quote:
Originally Posted by stratology View Post
OK, that's good practice.



Two different issues here - your bug report and good relating information, and bad general advice about system maintenance.


It's not about being 'comfortable', it's about messing up the functionality of your system. Again, when Spotlight indexing is disabled, LPX will not work correctly because it uses Spotlight. If a network device is not in use, it does not affect performance in any way. Things like 'System cleaning' in Onyx are unspecific, and completely unnecessary. If you do things like log 'cleaning' in Onyx, your sysdiagnose report will not have the relevant information that you deleted. Logs are useful, and in the rare case that something goes wrong and a runaway error creates large log files, the better option is to delete the affected log manually.
I merely suggested ONYX for an easy way of disabling indexing at a low level. It also has a simple tool for clearing the driver caches if you do disable one to prevent it from loading.

Logic uses Spotlight for its library features. So yes, disabling Spotlight indexing can affect that. I don’t use the library function so that warning was overlooked on my side. You can easily bypass its use in Logic by manually navigating to things in finder. Pain, however what is more painful glitches while tracking or dragging items from finder? The system safe approach would be turning off the various Search Results check boxes in System Preferences under Spotlight. That is user level control. But you lose the integrated OS X search ability for whatever you uncheck.

Sorry, network traffic makes a difference because most events other than the battery driver causing coreaudio issues are iCloud or push service launch daemons and agents. OS X thrives on inter communication between devices. Everything wants to talk to each other and report states. These low level events depending on how they were coded can take longer than a low coreaudio buffer can wait. Which causes the overloads. Other culprits would definitely be any system bus monitoring apps like SMC Fan Control. Anything that falls into the IOKit realm can cause trouble. Nature of the design. OS X is great at threading things that don’t need low latency. When it is time critical then you get into issues. It’s been like this since the beginning. That is why they are trying to implement more into the Grand Central Dispatch event scheduling system.
Old 29th September 2018
  #20
Lives for gear
 
stratology's Avatar
 

Summary: fundamental misunderstanding of how Spotlight works, of what caches do, of how memory management works, of how user accounts work (network locations are system wide, not user specific), of how kexts load and unload, etc.

When you use crap apps like Onyx or SMC fan control, you compromise your system and your hardware.

Bad advice based on lack of understanding.
Old 29th September 2018
  #21
Gear Head
Quote:
Originally Posted by stratology View Post
OK, that's good practice.



Two different issues here - your bug report and good relating information, and bad general advice about system maintenance.


It's not about being 'comfortable', it's about messing up the functionality of your system. Again, when Spotlight indexing is disabled, LPX will not work correctly because it uses Spotlight. If a network device is not in use, it does not affect performance in any way. Things like 'System cleaning' in Onyx are unspecific, and completely unnecessary. If you do things like log 'cleaning' in Onyx, your sysdiagnose report will not have the relevant information that you deleted. Logs are useful, and in the rare case that something goes wrong and a runaway error creates large log files, the better option is to delete the affected log manually.
Look, since you are concerned of other users potentially messing up their system, I’ll leave it at the original disclaimer. You’re hacking the system. You can f*#k it up if you are careless. This is not advice to follow blindly, it is just my findings. Repeat at you’re own risk. If in doubt, turn around and leave your system alone. Even if you’re ‘comfortable’ in these things, back up first, have a contingency in case you need to wipe and start over. **** happens. I didn’t walk through this unscathed but I learned as I went along. Yes, you can corrupt your system, and fixes are not always simple and easy.
Old 29th September 2018
  #22
Gear Head
Quote:
Originally Posted by stratology View Post
Summary: fundamental misunderstanding of how Spotlight works, of what caches do, of how memory management works, of how user accounts work (network locations are system wide, not user specific), of how kexts load and unload, etc.

When you use crap apps like Onyx or SMC fan control, you compromise your system and your hardware.

Bad advice based on lack of understanding.
You keep picking and choosing terms and reconstructing what I said to meet your argument. The troll is strong in you. I’ve said my peace. Other users are grown up enough to understand common sense without insulting someone’s efforts to help the community.
Old 29th September 2018
  #23
Lives for gear
 
stratology's Avatar
 

When someone suggests to others to f* up Logic by disabling Spotlight, calling him out on it is not insulting, and it's not personal. It's providing factual information on bad advice.
Old 29th September 2018
  #24
Gear Head
Quote:
Originally Posted by 8SlaiN8 View Post
I've tried it with zoom tac-2r (just bought it to test thunderbolt performance) and apart from gaining some more speed in roundtrip performance is almost the same, can't really tell the difference, very rare stutters... also, looking forward maybe newest logic 10.4.2 will fix smth... anyway, looking forward to to your report about uad tb3 interface
Logic Pro 10.4 Release Notes Lots of good fixes and features

The Logic Support team is awesome. I worked with Senior Specialist Chris L. He called me several times over a couple weeks to discuss and do screen shares to demonstrate the issue, ultimately passing along the battery bug info to the programming team for review. With a new OS out I am sure they will be working hard to get Logic rock solid. It is a major selling draw for their hardware.
Old 29th September 2018
  #25
Here for the gear
Quote:
Originally Posted by zenwolf View Post
Logic Pro 10.4 Release Notes Lots of good fixes and features

The Logic Support team is awesome. I worked with Senior Specialist Chris L. He called me several times over a couple weeks to discuss and do screen shares to demonstrate the issue, ultimately passing along the battery bug info to the programming team for review. With a new OS out I am sure they will be working hard to get Logic rock solid. It is a major selling draw for their hardware.
LOGIC FTW!!! :D
Old 1st October 2018
  #26
Lives for gear
 
maxwelldub's Avatar
 

on the phone with apple now trying to get this escalated, do you have a case number?
Old 1st October 2018
  #27
Lives for gear
 
stratology's Avatar
 

@zenwolf

I can not reproduce the issue on a MacBook Pro mid-2010, 2.66GHz i7, 8GB RAM, macOS Sierra 10.12.6, Logic Pro X 10.4.2. No spikes, no audio drop outs, no increased CPU usage once a minute observable - with 'Bowed Metal Space' in Alchemy and Chromaverb, same as in the video.

The AppleSmartBatteryManager.kext is version 161.0.0, the same version that can be seen in the plist in your video. So it looks like the kext was not updated between Sierra and Mojave. When I look at Console, no 'Restarting Poll Type 4' messages show up, no Logic messages that show Core Audio overload either. Only 'SmartBattery: finished polling type 4' shows up.


Getting in touch with Logic support at Apple was the best thing you could do to find out what's going on.

You mentioned submitting bug reports at the Apple Developer site, usually bugs are submitted via Product Feedback - Apple - they read all bug reports that come in this way, but you never get a response. Drawing public attention to a bug is not necessarily the best way to get it fixed - a single well written bug report, with a good description (like your video) can be much more effective than 20 poorly written bug reports.
Old 1st October 2018
  #28
Lives for gear
 
maxwelldub's Avatar
 

Spent 2 hours on the phone with Apple today. The problem is you do not want to speak to Logic support- this is a system wide OS issue and needed to be escalated to the Apple Software engineer team. I wasted a solid hour talking to Logic engineers who do not deal with CoreAudio issues.
Long story short I am told the Apple software team is aware of the issue and are working on it. I submitted some reporting info and when the subject line was 'T2 chip software errors' (filled out by Apple not me) or something similar, which confirmed my suspicions this is related to the T2 chip and seems to be confined to the 2018 MacBook Pros. I asked to swap my MacBook Pro for a working Mac and was told this isn't happening with iMacs and iMac Pros. I'm going to give them a week to fix then swap my MacBook Pro for an iMac. We'll see how far I get. I have the number of a senior Apple advisor and will update when I hear more.

Last edited by maxwelldub; 1st October 2018 at 11:54 PM..
Old 2nd October 2018
  #29
Gear Head
Quote:
Originally Posted by stratology View Post
@zenwolf

I can not reproduce the issue on a MacBook Pro mid-2010, 2.66GHz i7, 8GB RAM, macOS Sierra 10.12.6, Logic Pro X 10.4.2. No spikes, no audio drop outs, no increased CPU usage once a minute observable - with 'Bowed Metal Space' in Alchemy and Chromaverb, same as in the video.

The AppleSmartBatteryManager.kext is version 161.0.0, the same version that can be seen in the plist in your video. So it looks like the kext was not updated between Sierra and Mojave. When I look at Console, no 'Restarting Poll Type 4' messages show up, no Logic messages that show Core Audio overload either. Only 'SmartBattery: finished polling type 4' shows up.


Getting in touch with Logic support at Apple was the best thing you could do to find out what's going on.

You mentioned submitting bug reports at the Apple Developer site, usually bugs are submitted via Product Feedback - Apple - they read all bug reports that come in this way, but you never get a response. Drawing public attention to a bug is not necessarily the best way to get it fixed - a single well written bug report, with a good description (like your video) can be much more effective than 20 poorly written bug reports.
I appreciate you checking on your MBP. I haven’t installed the release version of Mojave. Still waiting on support from some audio developers, but it gives hope that a fix may have been pushed through. I’ll be sure to also submit a report through the Apple Feedback portal in the future. I have a developer account so I documented it through that and got a response for additional system reports during the glitch I was experiencing. Logic support was also helpful as I mentioned.
Old 2nd October 2018
  #30
Gear Head
Quote:
Originally Posted by maxwelldub View Post
on the phone with apple now trying to get this escalated, do you have a case number?
FWIW, I got a similar response after working with my support specialist. He just escalated but said it would most likely go to the OS system team. My dev bug report is 43737017. I mirrored it on Open Radar.
Topic:
Post Reply

Welcome to the Gearslutz Pro Audio Community!

Registration benefits include:
  • The ability to reply to and create new discussions
  • Access to members-only giveaways & competitions
  • Interact with VIP industry experts in our guest Q&As
  • Access to members-only sub forum discussions
  • Access to members-only Chat Room
  • Get INSTANT ACCESS to the world's best private pro audio Classifieds for only USD $20/year
  • Promote your eBay auctions and Reverb.com listings for free
  • Remove this message!
You need an account to post a reply. Create a username and password below and an account will be created and your post entered.


 
 
Slide to join now Processing…
Thread Tools
Search this Thread
Search this Thread:

Advanced Search
Forum Jump
Forum Jump