Solved by 1440p HiDPI: MacBook Pro 16“ is hot and noisy with an external monitor
axu2.medium.com
axu2.medium.com
system_profiler SPDisplaysDataType | grep -C1 Resolution
DELL P2715Q:
Resolution: 6016 x 3384
UI Looks like: 3008 x 1692 @ 60 Hz
I recommend displayplacer [0] for changing and automating macOS display settings. It lets you set all sorts of configs not available in the GUI. Should be easy to have displayplacer enable HiDPI on a 1440p monitor. And it doesn't make persistent system changes or require disabling SIP (i think?) like that script referenced in the article.
p.s. check out my pull request #62 for a bug fix! Funnily, it's for a scaling/HiDPI related issue. (though only matters when you don't use "mode: <>")
S32D850:
Resolution: 5120 x 2880 (5K/UHD+ - Ultra High Definition Plus)
UI Looks like: 2560 x 1440 @ 60.00HzI'd love a solution that doesn't require closing the lid, as well.
Let’s say you have a 1440p monitor running at native resolution without HiDPI enable.
Run this to see the (huge) list of available modes:
displayplacer list
Now run this with those values copied from above to change to the HiDPI mode:
displayplacer "id:<id> mode:<mode number>"Given that the linked script just seems to create some sort of XML file in the right directory, maybe this is doable directly? I made an issue in the displayplacer Github to this effect: https://github.com/jakehilborn/displayplacer/issues/69
Display Stream Compression, working perfectly in Catalina, is completely non functional in Big Sur. I don't mean that dramatically, it has not worked since public beta 1, and still doesn't.
Apple just doesn't care if you don't have a Pro Display XDR.
I’ve spent $5k on maxed out machines for my engineers, and it’s been a complete mess. Half of them have had to switch to a Windows machine as their primary so they can work without staring at a compiler spinning its wheels.
Maybe an Mx Pro will live up to the legacy of the 2012 era, but Apple’s obsession with thinness and mediocre cooling doesn’t give me much hope for pro users.
Why not Linux?
I’d rather config WINE to run a Windows app then configure and play with WSL. It’s a great tactic but Linux DE’s are so polished these days I think WSL is a poor choice.
WSL1 was not well designed. If you only worked with WSL1, you would be surprised how well WSL matured.
Anecdotally, my outdated i5 quad core is much snappier than the mbp i7 8 core because it can go all out at 69 watts/max boost for hours and not throttle for a second.
Throttling occurs when these values go over 100. Throttling is not directly linked to temperature. There is some kind back-off algorithm that means that even if you trip this thermal throttling threshold, it will take some time before throttling is disabled. Even if you CPU returns to a normal temperature.
Also, disable turbo boost.
I never did proper benchmarking, but my feeling was that resolution played a role. When using 1080p things were fast, but the resolution is unusable on a 27" 4k display. When using 2160p things were fast, but too small for my eyes. Any resolution in between (this implies things are being scaled?) was sluggish.
Note that this is a 13" model, so the problem does not seem to be restricted to the 16" model that the author is talking about.
Has anyone spoken to Apple Support regarding this?
I did a clean install of Catalina and same issue persisted so I think I have a hardware fault.
Not really solved then. Everyone I know that has a 16” has the same issue.
Since the top of the line model is the one most chosen by professionals, very few will not connect to an external monitor. Apple should be issuing refunds left and right after this fiasco.
I have felt like the performance isn't great, but I chalked it up to Docker and the Linux VM it uses. However, I also noticed that the computer was warm to the touch with fans running even when I hadn't used it for hours. I've taken to manually forcing it to sleep at the end of the day.
I'm excited to try this HiDPI hack, because it may solve the power usage problem and the fuzzy display. It would be a real shame to lose the built-in display if this only works in clamshell mode, but it just might be worth it.
Edit: I just realized that I can't use Touch ID in clamshell mode. That makes this a harder call.
It seems that Radeon is taking a lot of power even on idle screen and with low load. This increases temperature which in turn creates environment where throttling on CPU happens so not only Macbook gets hot but also it’s much slower.
After research I found out that I can either work in clamshell mode (and I’d rather grew fond of having MBP + external monitor) or find a way to mitigate it.
I decided to opt for eGPU which solved issue immediately. MacOS feel snappier, I don’t remember last time when fans started and it’s cool to touch. Not an exactly cheap solution for a i9 MBP but, eh, you do what you have to do.
eGPU brings its own share of problems to MacOS (like ejecting it is not a simple thing - I always have to resort to doing this manually from console) and pulling the cable will kernel panic, but the upside is that I get quite nice connector hub and photo editing software works much nicer.
I do think it’s a shame that Apple even designed MBP like that. You buy beefed out MBP (5600M wasn’t even available when I was buying it - and supposedly it fixes this issue) and it can’t handle so complex task as displaying the second monitor without launching the rocket engine.
No I don't.
Are you an iOS dev?
One is a BlackMagic eGPU Pro (since discontinued).
The main reason that I disconnected them, is that I plan to get an M1X/M2 laptop, once they come out, and I know that the new chips don't work with eGPUs, so I might as well get used to it (Makes for simpler under-desk wiring, too).
Also, the eGPUs seemed to introduce random crashes, and I don't like running on discontinued kit.
I'll have to have a chat with my lawyer, next time I see him, and see what would be a good way to deal with them.
The power/heat is due unnecessary clocking up when it isn't even doing any work. The non-dGPU version does better at the same task.
If that would be solely driver bug it would be patched by now (and that’s an issue since 2019, so not exactly new).
my Mac is asking me every login for my exchange creds despite having authenticated several times.
my Mac is running hot and depleting battery without any discernable reason at random occasions.
the esc key on my mbp 2019 itself is a bug in my opinion.
I have to disagree there. It's so frigging laggy that it feels unusable as a communication tool.
Just yesterday we spent time in a team meeting because someones Teams instance was malfunctioning.
Sometimes restarting the application completely (be aware it idles in the systray) works, other times restarting the pc or disabling (or enabling) hardware acceleration will work. Other times it helps to move it to the built in monitor it seems.
And sometimes, like yesterday, nothing including
- all the above,
- disabling incoming video
- and reinstalling the application
works.
As for UX issues there's clicking on "join meeting" and continue working for a few minutes until you realize "ah, the first join meeting doesn't join the meeting, only opens the join meeting dialog".
Another winner: when you click on the 7 more people icon to see who else are online and it does nothing.
Or the fact that joining a started meeting is a multi step process (find calendar, find today, find the meeting click join and join).
Yesterday I'm also sure I had an instance of sound getting through before I joined.
Still, somethings works flawlessly for me: jumping seamlessly from one device to another during a meeting is fantastic.
Sound is OK for me. Background switching is ok etc etc.
- Chromium: infinite login loop because why not
- Flatpak: login works, but says my account isn't eligible for Teams (which blocks joining as even a guest?)
- Firefox: doesn't even ask for a login and I can join, but no audio or video (either direction); chat works though
There's an authentication bug that causes a log in loop for anyone who signed up for a personal account using an email address that they also used as part of another MSN/live.com service (eg- as a secondary address), especially if your original Teams/O365 account was converted to a domain/business account.
If you log into live.com with your Teams email, add a different fallback email and remove your current one, that may help.
They missed the entire container revolution with docker. I still come across Devs using Mac's that are afraid of docker because it's too confusing and black box. (It's Linux in there right?)
At work we have a rather overengineered method of proxying to our production services for security reasons. Mac users are currently constantly having to deal with abstraction layers on top of abstraction layers to make things barely reliable. I just use a systemd unit file and haven't looked at it in years.
Homebrew tool, while great for more obscure things, it should really only be a fallback, not the default. It's basically a confusing and black box version of the AUR.
BSD/apple ways of doing things are just annoying. It's fine for the average user. But for Devs that want to do things in the same way they do them on their server it's just another hoop to jump through.
The sad thing is that moving every Dev in the company over to Linux would probably be worthwhile long term, but I really don't think they have the willpower to relearn things even if they are that much better.
The issues are 100% due to Apple's lack of care for developer concerns.
I've never used Outlook for Mac, but there's the Apple Mail.app on the mac which works great with Exchange. I'd choose it over Outlook on Windows every time.
There is a caldav addon that allows syncing with remote calendars. I use that to coordinate calendars on my computers and iphone.
Thunderbird had some big hi changes over that time, you might like it.
I dunno, I'm a developer (Python, Java, some JS, shell, Docker, Vagrant) and it has worked fine for me.
Also it was as easy as macOS install: I just installed it in a 5 minutes and it was ready to work. With Windows I had to spent hours to install all the drivers from the manufacturer website (I tried to let Windows to install their own drivers, but it did not find all the necessary drivers).
I have few issues with some Linux UI choices, for example I don't like Gnome 3 UI. But that's not a bug.
The only thing I've had to do after installation, aside from installing my apps from repos, is to add the Mainline PPA and update the kernel to newest just to take advantage of newer kernel drivers for some of my hardware. Other than that, everything is there and everything works.
PS I just tried to use it. I've got some permission denied errors from containers which were not present with docker. I guess podman is not ready as a complete replacement of docker for me.
it's not ready for replacement.
If I would write containers myself from the scratch, I'd definitely try to ensure that podman works. But I'm using containers from other people, so I have to live with that for now.
sudo dnf install -y podman-docker
cat /usr/bin/docker
#!/bin/sh
[ -f /etc/containers/nodocker ] || \
echo "Emulate Docker CLI using podman. Create /etc/containers/nodocker to quiet msg." >&2
exec /usr/bin/podman "$@" echo "user.max_user_namespaces=28633" | sudo tee -a /etc/sysctl.d/98-userns.conf
sudo sysctl --system
sudo usermod -v 10000-65536 -w 10000-65536 $(id -un)
podman system migrate
podman unshare cat /proc/self/uid_map
podman infoYour "average" Linux distro isn't "buggy" as you claim.
You sound like someone who either hasn't used Linux, or extrapolates their experiences from one specific distro, to all distros.
I say this as someone who uses Linux daily. Specifically, Arch Linux, with testing repos enabled. The only bugs I run into are from using Git master versions of software that I'm interesting in testing.
Linux does exceptionally well considering there is almost no money at all going into this use case. But there are many bugs, sometimes due to hardware (e.g ACPI inconsistencies, even Thinkpads have issues) and because of software churn (few fix bugs in their spare time, it's more fun to create something new).
If you wish to pretend that everything is perfect and there are no problems, then you aren't helping Linux either.
This is me figuring out my Optimus without Mux. My screen freezes when laptops (ASUS N76VB) resumes from sleep. (I don't know which state S1 - S4.)
So I have "Fixed" the issue. But this was a problem of the HD Audio device driver? in the NVIDIA 740m?
It was difficult to fix and I'm thinking about making a GitHub repository thing about it.
I had this issue on Fedora, Ubuntu and Pop!_OS.
Desktop UI is where Linux tends to shine. Laptops... oh, those are a royal mess, usually.
> Linux does exceptionally well considering there is almost no money at all going into this use case.
No money at all? There are countless developers being paid to work on the Linux kernel. And on other pieces of software in userspace ~ usually by Red Hat.
> But there are many bugs, sometimes due to hardware (e.g ACPI inconsistencies, even Thinkpads have issues)
Hardware-related bugs are always an unpleasant can of fun...
> and because of software churn (few fix bugs in their spare time, it's more fun to create something new).
Bugs caused by software churn are also a thing. I personally haven't noticed anything major from many packages in the main repos.
> If you wish to pretend that everything is perfect and there are no problems, then you aren't helping Linux either.
Oops, I could have worded things slightly better, as in "not as "buggy"' instead of 'not "buggy"'. I sometimes miss words, only to realize later.
But I'm not saying this as blaming Linux about the brokedness. They're doing amazingly well for the huge amount of combinations of hardware and software they're supporting.
Apple, on the other hand, has decided to seriously limit the amount of hardware they support, which makes it a much easier effort to support those ones.
Ironically you've just described the biggest problem with Linux, and the biggest non-solution often proposed "Problem with distro Y? Just use distro X".
Whenever a solid Apple bash comes up on HN, people just throw statements around to the effect of "Apple -- as an entire software and hardware conglomerate -- just doesn't 'care' any more," and that nothing in the entire freaking stack should ever exhibit a bug. It's almost like NO ONE around here actually works in SOFTWARE or HARDWARE. I guess no one else here has ever shipped an uncaught bug. I sure have.
From my experience, the number of the kinds of issues being discussed here is EXCEEDINGLY low for the number of moving parts involved. I know, I know. Someone's going to say that since they do the whole stack themselves, they should NEVER ship a bug. Well, good luck with that. I'll believe it's humanly possible when I see it.
And don't even start with how Linux or Windows is any better. I've been at this for 30 years. I know better.
Here we are at 11.3.1 and it is still absolutely, utterly broken. Legions of users of varying monitors, varying GPUs, can't access the full feature set of their equipment that they used to be able. There are Radar reports back to the betas.
Instead we get things like making sure you can buy AppleCare+ from the About This Mac option.
Or buy a Pro Display XDR, because it seems that that is all Apple cares about for external monitors.
The monitor issue was particularly egregious because it would cause heat and noise even at idle, but the thing gets perilously hot and loud when put under even the slightest load.
I am incredibly surprised that none of the reviews seemed to pick up on it, though I suspect they may have been highly swayed by this being the first MacBook in some years with a usable keyboard.
I would not be surprised if we start to see a bunch of them fail en masse in a few years time.
This sentence perfectly encapsulates the Apple mentality.
Why would they ever bother fixing it?
I noticed the issue as my laptop feeling laggy overall, but especially that graphics on the laptop screen (eg the minimise animation) ran at like 5fps when my external monitor was connected - they were fine without, and the external monitor was ok.
A friend said theirs was fine with the same setup but a 4K monitor so I did wonder if it was something to do with scaling but wasn’t aware you could do anything about it. Will test this out tomorrow!
Slightly off topic but related - had a fun day yesterday tracking down the fact that in Big Sur, timers in tasks that are not in the foreground (ie the active window) start firing up to 10 secs late. I thought it must be a symptom of whatever issue was making my computer laggy but it actually seems to be aggressive power saving... so if anyone knows how to opt out of that I’ll be happy, really messes up what I was doing yesterday (basically a program that forwards IPC messages between two programs - they start arriving late after 10 secs or so).
Run `sudo watch -d -t -n1 thermal levels`.
When levels are above 100, your laptop will throttle. Once they return to 100, throttling will cease. There is a back-off algorithm at play too, so its not good enough that your CPU returns to normal temperature, it must stay at normal temperature for a period of time before things stop lagging.
Also, disable turbo boost.
Is there a way to disable it without installing some 3rd party software?
If I force internal GPU with `sudo pmset -b gpuswitch 0`, animations and graphics benchmarks are much faster than with dGPU enabled! For example, I tried the Unigine Heaven benchmark and I get 14fps with internal GPU, 4fps with discrete (5300M) GPU, which I am pretty sure is not right lol. (If you run this command, `sudo pmset -b gpuswitch 2` set it back to auto).
Did a clean install of Catalina to rule out OS issues, and got friends with the same laptop to test out the same website I was seeing extreme slowness on (it was fine for them).
The laptop recently had a repair and pretty much every part was replaced as a result of liquid damage, so I suspect either a faulty part or incorrect installation. Off to the Apple Store I go!
This issue has been plaguing me for ages and I only just now figured out the solution!
The biggest obstacle to blogging is vacillating endlessly about the perfect place to post it. Post wherever it's easiest.
(none of the people who reply to this comment will actually blog regularly; they just like to give advice)
I assume everything on Medium is professionally stupid, based on what I've read there so far.
Substack seems pretty popular and not terrible.
Edit: Apparently a bad idea: Dev.to seems to be fairly decent, and maybe more like medium.
Hosting your own static site via netlify (or similar services) is only the cost of a domain name.
Hosting your own vps-based site on gcp is also only the price of a domain name, and if you put the also free cloudflare in front of it, it scales well, but this route does mean a lot more maintenance than any of the options above.
Are DOS attacks that cause lots of outgoing bandwidth common? And wouldn't launching such a DOS attack be somewhat expensive in the first place?
Also, I don't think any of the hosting companies I use have metered bandwidth. I assume that if I used excessive bandwidth I'd just end up saturating the network, but for a blog with mostly text and a few images that shouldn't be an issue.
It's any traffic. It can be invalid packets if that's effective, but it can also be complete http requests.
> Are DOS attacks that cause lots of outgoing bandwidth common?
They're commonly about saturating the incoming bandwidth. But there's nothing wrong with more outgoing traffic if someone wants to hit you that way.
> And wouldn't launching such a DOS attack be somewhat expensive in the first place?
Not if you're using other people's hacked endpoints. (Which is pretty much the default these days)
If you use a service with just a data cap, that's fine. I commented with AWS / GCP / others in mind.
It's perfectly normal for legitimate traffic to accidently DDoS you if you have low specs (or a huge unwieldy site) and get a link somewhere popular like HN.
I run a blog on a small cloud VM. If it made front page of HN all those visitors would DDoS my site "by accident" in no time.
Don't forget, accessing a website the normal way is outbound bandwidth.
Incidentally, I like to use my blog to explain topics that aren't widely understood outside of tech people or domain specific experience.
Maybe I should be a bit more active and explain things like this?
Besides, it lowers latency for users by being a real cdn with a lot of points of presence, and therefore improves the experience.
My bet is that another scaling step is causing it; and HiDPI mode is "native", so no scaling.
You can see this in the comparison of the two screenshots; the first one looks like an antialising filter was applied to all the pixels and halved the resolution (so a 2x2 block of 4 pixels turns into one averaged "huge" pixel), the second one doesn't.
That's not always the case. It can be, depending on the resolution selected (in this case it's "default for display" so no idea).
HiDPI just means 2x rendering. But it still might be scaled to the resolution you've chosen for your display.
On my MacBook Pro 13", there are four scale choices from "larger text" to "more space". The first option upscales (slightly blurry), the second option displays the 2x directly ("natively"), and the third and fourth options downscale.
They're all still HiDPI though. In contrast, when I plug into my 1080p projector, it goes back to 1x ("LoDPI"?) rendering.
macOS has two rendering modes: Normal (1x) and HiDPI / Retina (2x). For this reason, most apps have two different sets of assets like icons and other images shown in the UI.
After rendering the content either at 1x or 2x, the picture is scaled to the "real" resolution of the display. This allows Apple to support a large number of "scaled" resolutions, but the apps only have to support two modes (1x or 2x).
You can see that this happens if you have two displays attached to your Mac, one with Retina and one without. When you drag a window from the Non-Retina-Display to the Retina-Display it will be pixelated at first (it was drawin in 1x mode), until it's completely on the new display, then it will redraw in 2x mode and become sharp.
Normally 1x mode is faster. Try opening an app that shows a lot of data (eg. a large spreadsheet) and scroll quickly on a 1080p display vs. on a 4K display. You'll should see that the scrolling is much smoother on 1x than on 2x, since the 2x display needs to draw four times as many pixels.
So this is clearly a bug, and not because of some additional scaling from 2x->1x. As this blog post shows, it's faster when downscaling from 2x.
(Also, the pictures in the blog look like they have been resized for the blog, so I wouldn't read too much into their blurriness)
https://en.wikipedia.org/wiki/Intel_Graphics_Technology#Skyl...
The last Macbook Pro with Iris Pro graphics is the Mid-2015 15-inch Macbook Pro.
https://en.wikipedia.org/wiki/MacBook_Pro#Technical_specific...
The Mac in the article in particular is a 2019 16-inch model.
Too bad it can’t be used with external monitors. It’s probably quite power efficient.
Intel GPUs (including the UHD Graphics 630) can be used with external monitors. HP, Dell, Lenovo, Acer etc all do it.
In fact, in 2013-2015, Macbook Pros were available with ONLY Intel graphics, and supported external monitors.
The reason why you cannot do it with dual-GPU Macbook Pros is that Apple has their own display muxing solution that requires the discrete GPU to be powered up to use external monitors (presumably the ports are attached to the discrete GPU).
I am not sure about this part:
>y MBP is well behaved when I dont plug in an external monitor, and noisy/hot when I do. This is with both OSX and Linux.
Because it doesn't on my laptops and screen.
They really need to put some focus into it because Windows really makes macOS look bad in this regard.
For the past year I've been WFH with the MBP (which has a 5500M gpu) driving its own panel (3072x1920), an external "wide" ~4K~ "Big-ass" (3440 x 1440) and an external "4:3" 4k (3840x2160)
No problems here.
[edit: display info]
tanelorn ~ % system_profiler SPDisplaysDataType | grep -C1 Resolution
Display Type: Built-In Retina LCD
Resolution: 3072 x 1920 Retina
Framebuffer Depth: 24-Bit Color (ARGB8888)
--
V3L6W:
Resolution: 3440 x 1440 (UWQHD - Ultra-Wide Quad HD)
UI Looks like: 3440 x 1440 @ 60 Hz
--
PN-K321:
Resolution: 3840 x 2160 (2160p/4K UHD 1 - Ultra High Definition)
UI Looks like: 3840 x 2160 @ 30 HzSome of the issues include: screens refusing to turn on, random screen shutdowns (then turns back on), incorrect refresh rate on wake from sleep, incorrect resolution on wake from sleep, monitor position gets reset again on wake, screen image offset on monitor, once detected resolutions and capabilities not detected after restart (hdr, 4k, brightness control), refusing to pass audio through, clamshell mode will keep internal display active…
I’m not alone (and this isn’t limited to M1)
"screens refusing to turn on, random screen shutdowns (then turns back on), incorrect refresh rate on wake from sleep, incorrect resolution on wake from sleep, monitor position gets reset again on wake, screen image offset on monitor, once detected resolutions and capabilities not detected after restart (hdr, 4k, brightness control)"
And this can be the display itself that's causing the issue.
An example that i've found:
```If you suspect that you are affected by this issue, see if you find a reference to YCbCr in the menu of the monitor settings. For the Dell, it was in > Menu > Color > Input Color Format``` Dell (P2720DC)
The funny thing is running Windows on my Mac mini via bootcamp shows none of these problems.
And it's what I expected. MacOS or the Display manufacturers treat each other differently in terms of information shared between the display and device itself.
I'm talking about EDID.
https://gist.github.com/ejdyksen/8302862
Not every display, I mean device is built the same. Even if it's the same manufacturer or same device.
That is so annoying on my machine, a band an inch wide is offset by an eight of an inch when it comes out of sleep one time in five. The only solution is a reboot.
I've had all the other problems as well, on both Intel and M1.
In Catalina my Mac Pro with a W5700X happily drove 2 4K monitors in HDR10 at 165Hz.
In Big Sur my options are now SDR at 95Hz or HDR10 at 60Hz.
This has been broken since the 11.0 betas. This happens with multiple GPUs, multiple monitors, multiple interface types.
I originally came to the MBP from the world of windows and being very jaded by BIOS and sleep issues with my Dell XPS 15. Macs just work they said. Well, jokes been on me. I literally have to constantly mute myself during video calls because my $3500 MBP fans are always whirring away so loudly that it ends up fucking up the audio for others during a call. I've had my entire system throttle and grind to a halt during video calls because of thermal throttling when temperatures got higher during this summer.
I'm so over Apple and their laptops. My wife's MBP has a swollen battery that's destroyed the touchbar and touchpad. For all my complaints with Dell, it was nothing in comparison to the pain of owning my MBP. Multimonitor support is a joke. Docks are a joke. Opening apps is a fucking joke due to their sandbox slowing things down to the point that opening any app takes 5-10 seconds.
At this point all I do is use my $3500 MBP as a VNC client to my Linux desktop where I get all my real work done. This is the first and last Mac I will personally own. I can see why they were as popular as they were with the developer crowd. But Apple has lost its way and doesn't give a shit, at least not when it comes to the developer / professional users that they originally targeted with their MBP line.
I'll be going back to Windows/Linux soon and there is very little I will miss when I make the switch.
This is pretty much by sentiment too. My MBP is a "2nd" computer for me, that I only use when building iOS apps - my daily driver is Windows 10, and I'm very happy with it.
It's not quite 50-50 for kernel panic for me, but I'd guess it happens around 20% of the time if the MBP is switched on when I plug in the monitors - hardly much better. The keyboard on my MBP is crap too, and has 2 keys that stop working from time to time. I'm not even a fan of the trackpad that most on HN seem to gush about - it needs too much pressure to click even on the most sensitive setting, and something just feels "weird" about the feel of it and the way teh cursor moves. And it seems something breaks or goes wrong after every.single.os.update - IMO, there are some serious QA issues with OS updates.
Meanwhile on Windows 10, everything Just Works (TM).
That being said, 8-9W with closed lid is still high. Consumption for second monitor when it's idle shouldn't be more than 1W.
After seeing the title I got my hopes up, that it will work within 1-2W range for idle regardless of lid being closed. That would be a solution, 8-9W definitely is not.
I too looked at the title and thought someone had finally figured out how to get around having the laptop open and a monitor plugged in with no noise, but it seems like this basically makes no difference at all since it only works when the lid is shut. I've resigned myself to the fact that it'll never be solved, and hope to get this MBP replaced at some point
Will switch to Apple Silicon version of 16 when it comes out, but boy does that leave bad taste in mouth.
After buying Mac during a 4 year run of devices with unfixably broken keyboards and GPUs, that will not teach apple to focus on quality or repairability.
Of course, because you can dissipate enough heat with the lid closed.
>That being said, 8-9W with closed lid is still high. Consumption for second monitor when it's idle shouldn't be more than 1W.
I am not about that, because laptop displays draw more power than 1W most of the time. (I'm aware of those low power displays that Lenovo utilizes.)
And there are a couple of factors that I've forgotten.
B326HK:
Resolution: 3840 x 2160 (2160p/4K UHD 1 - Ultra High Definition)
UI Looks like: 1920 x 1080 @ 30.00Hz
The M1s are a huge improvement but the bar was pretty low.
You would hope so, but behemoths like Apple just don't care and don't have to. I'm guessing it would take thousands of returns before it even made a blip on their radar.
Because they give all their devs 5k displays.
Edit: well we have uh fewer employees …
How were you able to return it? My MBP 16' gets really hot without an external monitor. And I'm getting nowhere with my complaints.
Though, a few months ago I had contacted Apple about a different matter and mentioned these problems to the agent who then told me that was the first time they and their colleagues had heard of this issue.
Their tone and the very many words they used to convey this told a different story. Going to be an interesting case when I finally have time to get this sorted.
Edit: Not saying it's company policy to deny this issue exists. It could just have been this agent not wanting to deal with it (maybe because it'd end up in a concession which would look bad on his profile?)
Apparently there are some AMD desktop GPUs that exhibit similar behaviour, but obviously an extra 15-20W draw is far less of an issue in a desktop than a thin laptop.
Either way, this machine is coming up on two years old. It’s not gonna be fixed, ever. I already sold mine because of it.
First, disable turbo boost.
Second, point a desk fan at your MacBook.
Third, eliminate processes that peg the CPU.
---
How to know when it is throttling:
Download Intel Power Gadget and watch the Frequency graph. If blue line doesn't match the red line then you are throttling.
Run `sudo watch -d -t -n1 thermal levels`. When any reading goes over 100, it is in throttling mode.
Once it trips over 100 you start throttling. Throttling doesn't end though until this value returns below 100. This number is not directly linked to temperature either. Even though your CPU may be at a safe temperature like 80 degrees which normally shouldn't throttle, your `thermal levels` will not decrease immediately. The best way to get them to decrease is by sleeping your MacBook.
Maybe there is a hard-coded delayed back-off algorithm that is a function of the time the CPU spends above a certain temperature threshold - I'm not sure.
I have also read that its potentially the voltage-regulator module that doesn't have a temp sensor that this may be linked to.
---
You can also do a [cooling mod][1] if you really want to but its probably going to void and warranty and its risky.
[1]: https://forums.macrumors.com/threads/2018-macbook-pro-coolin...
I play an occasional Dota 2 match, that my 2014 13" handled without any issues, while my 16" just shuts itself down because of overheating (even while not docked, because of dGPU usage). I don't care if it's an AMD issue, then the GPUs shouldn't be in their machine.
MacBooks used to be so good, and just work. Now not so much. I'll think twice before buying anything Apple (computer-wise) ever again.
Specs: 2,3Ghz 8-core i9, 64GB RAM, AMD Radeon Pro 5500M 8GB
No doubt that's true; I haven't observed this particular problem. However, OSX' logic for external monitors doesn't work a whole lot better than for older laptops -- in particular, if I connect it to a monitor that's connected through a Startech KVM it remains stuck at XGA (1024x768) resolution, which reflects the EDID the KVM provides to inactive computers. It doesn't update when the laptop is selected, and there seems to be no way to reset it.
Nothing else does that. IMO, it's all indicative of a prototype-level implementation.
A lot of the time kernal_task would take up like 400% CPU, which was because plugging the power cable into certain thunderbolt ports (on the left side) would overheat the computer.
Even then I was consistently getting low performance, my fans were on all the time, and my computer was burning hot trying to power a 4k monitor and a 1080p monitor.
After my 16" broke I got an M1 Mac Mini and obviously have not been having the same issues. Now my computer feels fast again. My old set up felt like I was using a Macbook Air or something.
After some searching it seems HDMI conversion can trigger it. Switched to a TB4 hub that did usb-c to monitors (which uses display port instead, Kensington) and radeon high side idle went back down and fans slowed down.
Will be happy to move to Apple Silicon just for the lack of fan noise.
Is this by design?
Note: Mac lid is closed during all of the testing and usage I am doing.
Because they do things differently and the laptop that using is doing things differently too. And... about those other laptops, they do things differently too. (ASUS Zephyrus vs a Legion 7 differs a lot.)
Another data point. Seems like putting more load on some part of the system prevents this, but it is very strange.
The amount of defects and other miscellaneous issues I've encountered, along with numerous productivity losses moving from Windows and Linux. Makes no sense. My computer needs to function as a productivity machine. I don't have time for it.
Next laptop(s) are going to be ThinkPads. Because I can buy two for the price of a macbook. We program for Linux servers anyway.
"Note, these power draw settings only work if you keep the MacBook lid closed."
I am also not sure about the OP's statement. (I am not saying this person is wrong.)
I have always ran a 4K monitor and never hit this issue. My 15" (2018) and 16" MacBook Pro run silent when connected to the 4K monitor as it uses 2x scaling by default (so 4K "looks like" 1080p).
However if I connect a 1440p display after half an hour or so the fans kick in.
I don't own a 1440p display anymore to test this solution unfortunately.
> You can tell, since the screenshot of the power draw monitoring utility (iStat Menus) is twice as sharp in the second example.
Sorry, didn't realize paywall was default option: https://axu2.medium.com/solved-by-1440p-hidpi-macbook-pro-16...
Basically the tradeoff is to have better circulation among subscribers or non-subscribers. I always choose the latter.
Can we ban medium posts on HN already? imagine if we login-walled every piece on HN. we should not encourage this behavior and we have the clout to send Medium a message.
medium literally exists to just put a wall between you and the authors content. nobody's making real money blogging on medium. and HN is an audience of developers, for whom standing up their own blog is trivial.
ironically dev.to is shadowbanned on HN (understandable because lots of blogpspam), when medium is the one that gives readers a terrible reading experience.
Links should only be allowed to the public internet.
I didn't study the content closely, but it certainly looked like the flags were fair.
I'm not sure we do. Medium is comparatively a much bigger and more mainstream platform than HN.
Because I block all cookies by default, I never run into this problem.
It was great but it was also a dread, even after training it. If you visit mostly the same sites I guess it's fine.
It still works.
"It's ok to post stories from sites with paywalls that have workarounds. In comments, it's ok to ask how to read an article and to help other users do so. But please don't post complaints about paywalls. Those are off topic. More here."
I've been asking this for years.
Perhaps there's a need gap for 3rd party service which can work out a deal between news websites and submission forums.
I was flagged.
HN is carefully curated for our FAANG overlords.
As for the mods: we don't care about $BigCo or "overlords", but definitely draw the line at single-purpose accounts and serial trolls.