(2019) 16" is HOT & NOISY with an external monitor! :(

[kaans]

The device is truly a Pro machine on Windows, I had to spend 4-5 hours doing builds, installed Hyper-V, a lot of processing etc. - while I was waiting for stuff to complete/download/upload, downloaded the original Battlefront, gamed a bit

All this and the max temp was around 80C, even while gaming, the temps were 70C at max - on average/idle - 60C

TL;DR: Windows is 10C cooler than MacOS, practical difference is huge

It's just bizarre, the device feels safe to use on Windows, but everytime I got it connected to an External screen on MacOS, I make sure it's able to sleep so it doesn't burn itself off, and oftentimes, I disconnect the screen if there's a lot of processing to be done, such a hassle :/

 

[jaduffy007]

Hey people I just hopped over from the Apple discussions forum to this thread. Did you get anything useful or is there a page # or pages of particular interest?
Thanks, Peter

The SwitchResX "fix" is very worth knowing about...page 75, I believe.
[automerge]1593314629[/automerge]

I returned my 5500M model and have been using the 2.4GHz/32GB/5600M/2TB model for a few days now. No issues so far using a 1440p 165Hz external monitor. anyone know of an alternative to iStats? I cant get power consumption metrics with the 5600M from iStats.

Hmmm ... I didn't think high refresh rate monitors were having the same heat / fans issues as 60hz monitors.

 

[Grohowiak]

The SwitchResX "fix" is very worth knowing about...page 75, I believe.
[automerge]1593314629[/automerge]

Hmmm ... I didn't think high refresh rate monitors were having the same heat / fans issues as 60hz monitors.

I have the base model and the SwitchResX doesn't do anything with lid open and 2K external screen.
Radeon is still pushing 18w+. CPU seems normal depending workflow but it idles at around 5-6w.

 

[consoletuner]

Hi folks

Ive received answer from the creators of iStats. They provided me a beta of iStats.

Can please someone with the 5600M check the power draw:

https://cdn.bjango.com/files/istatmenus6/1135.zip

 

[Minga089]

I have the base model and the SwitchResX doesn't do anything with lid open and 2K external screen.
Radeon is still pushing 18w+. CPU seems normal depending workflow but it ideas at around 5-6w.

Same here. That "fix" if you want to call it like that works only for clamshell mode.

 

[alexktz]

Hi folks

Ive received answer from the creators of iStats. They provided me a beta of iStats.

Can please someone with the 5600M check the power draw:

https://cdn.bjango.com/files/istatmenus6/1135.zip

The BETA does indeed work, thank you!

8 watts on a Dell 34" U3419W ultrawide @ 3440x1440 connected via 1 usb-c cable (this monitor does not have thunderbolt). I just returned my 1 week old 5500m (after trying the SwitchResX stuff and failing) for a 5600m in the hopes that the HBM memory would resolve the heat, and it has. Exactly the same setup with the monitor between the two GPUs, same monitor with the 5500m was 20+w and *heat*.

This is the top i9, 32gb ram, 5600m spec.

1 hour connected to the external display and fans are yet to go above 2000 rpm during normal coding, web browsing, etc. If I compile something they obviously ramp up for a bit but that's totally fine.

HOORAY!!!!

 

[Minga089]

The BETA does indeed work, thank you!

8 watts on a Dell 34" U3419W ultrawide @ 3440x1440 connected via 1 usb-c cable (this monitor does not have thunderbolt). I just return my 1 week old 5500m for a 5600m in the hopes that the HBM memory would resolve the heat, and it has.

This is the top i9, 32gb ram, 5600m spec.

1 hour connected to the external display and fans are yet to go above 2000 rpm during normal coding, web browsing, etc. If I compile something they obviously ramp up for a bit but that's totally fine.

HOORAY!!!!

With open lid?

 

[alexktz]

With open lid?

Correct, yup. Open lid. I'm so happy!

 

[consoletuner]

The BETA does indeed work, thank you!

8 watts on a Dell 34" U3419W ultrawide @ 3440x1440 connected via 1 usb-c cable (this monitor does not have thunderbolt). I just returned my 1 week old 5500m (after trying the SwitchResX stuff and failing) for a 5600m in the hopes that the HBM memory would resolve the heat, and it has. Exactly the same setup with the monitor between the two GPUs, same monitor with the 5500m was 20+w and *heat*.

This is the top i9, 32gb ram, 5600m spec.

1 hour connected to the external display and fans are yet to go above 2000 rpm during normal coding, web browsing, etc. If I compile something they obviously ramp up for a bit but that's totally fine.

HOORAY!!!!

OMG! Amazing, thanks a lot for your time!!!

Then I know what to do!
Somehow Im happy now, but as well sad for the 5500M guys
[automerge]1593361903[/automerge]

Correct, yup. Open lid. I'm so happy!

Stupid question: With closed lid you have as well the low Wattage?

 

[alexktz]

OMG! Amazing, thanks a lot for your time!!!

Then I know what to do!
Somehow Im happy now, but as well sad for the 5500M guys
[automerge]1593361903[/automerge]


Stupid question: With closed lid you have as well the low Wattage?

Lid closed is 4-12w.

Lid open is 8-12w.

 

[alexktz]

Don't know how relevant this is but the two laptops are both running 'the most recent Catalina' available to them, 10.15.5 but their builds are different.

5500m - 19f101
5600m - 19f2200

So then I thought maybe these different OS builds have different Radeon drivers? I think that might be it, so perhaps there is hope these newer drivers will 'trickle down' to 5500m builds soon.

5500m

• EFI Driver Version - 01.01.190
• ROM Revision - 113-D3220E-190
• VIOS and Option ROM version - 113-D32206U1-019

5600m

• EFI Driver Version - 01.01.192
• ROM Revision - 113-D3000E-192
• VIOS and Option ROM version - 113-D3000A0U-014

Now I have zero idea what any of this means and it's likely that some of it is due to the fact it is a different GPU after all. But maybe this info is useful to someone.

 

[09872738]

Thx for all the info. Now, one question remains:

What if you attach 2 high res monitors with lid open? Anyone?

 

[interbear]

The BETA does indeed work, thank you!

8 watts on a Dell 34" U3419W ultrawide @ 3440x1440 connected via 1 usb-c cable (this monitor does not have thunderbolt). I just returned my 1 week old 5500m (after trying the SwitchResX stuff and failing) for a 5600m in the hopes that the HBM memory would resolve the heat, and it has. Exactly the same setup with the monitor between the two GPUs, same monitor with the 5500m was 20+w and *heat*.

This is the top i9, 32gb ram, 5600m spec.

1 hour connected to the external display and fans are yet to go above 2000 rpm during normal coding, web browsing, etc. If I compile something they obviously ramp up for a bit but that's totally fine.

HOORAY!!!!

This is very interesting.

If I could be assured of quiet fans when connected, lid open, to a monitor, I’d pay the difference and upgrade to this model if I could.

 

[Norbert Mikołajczyk]

That looks like what I'm seeing with my 5500M computer. I also don't see the "max fans as soon as monitor is connected" issue that most are seeing despite the GPU jumping to 20W immediately.

Looks like they haven't fixed the issue with the 5600M after all. And all indications are pointing to it being a driver issue as well at this point.

Yeah, Mac drivers are garbage in comparison to Windows...

Don't know how relevant this is but the two laptops are both running 'the most recent Catalina' available to them, 10.15.5 but their builds are different.

5500m - 19f101
5600m - 19f2200

So then I thought maybe these different OS builds have different Radeon drivers? I think that might be it, so perhaps there is hope these newer drivers will 'trickle down' to 5500m builds soon.

5500m

• EFI Driver Version - 01.01.190
• ROM Revision - 113-D3220E-190
• VIOS and Option ROM version - 113-D32206U1-019

5600m

• EFI Driver Version - 01.01.192
• ROM Revision - 113-D3000E-192
• VIOS and Option ROM version - 113-D3000A0U-014

Now I have zero idea what any of this means and it's likely that some of it is due to the fact it is a different GPU after all. But maybe this info is useful to someone.

Here is the link to that update, I wonder if 5500m users could install it and update the drivers. I have poor internet connection right now to check it myself: https://support.apple.com/kb/DL2044

 

[HowardEv]

I received a base 16" today and am disappointed because it get's as hot as my 2017 with an external 4K monitor.

Is anybody else experiencing the same issue? I always work with external monitors and running on high resolution and this heating performance is upsetting. It's not cooler than the 2017 version in this respects.

Can you guys share your thoughts on this?

Thanks a lot!

I recommend waiting for the "new form factor" ARM iMacs there was a story about here last week. The new form factor should address the heat problems by splitting a MacBook in to two parts: the screen and keyboard, and a stationary base with a proper heat sink that connects wirelessly to screens and laptops using that new wireless hi-res technology invented for the goggles prototype. Old 4K monitors could still connect with a USB C, and charge it too.

 

[jaduffy007]

I recommend waiting for the "new form factor" ARM iMacs there was a story about here last week. The new form factor should address the heat problems by splitting a MacBook in to two parts: the screen and keyboard, and a stationary base with a proper heat sink that connects wirelessly to screens and laptops using that new wireless hi-res technology invented for the goggles prototype. Old 4K monitors could still connect with a USB C, and charge it too.

That's exactly what I intend to do...replace my 16" with an ARM iMac + ARM entry level 13" (14") MBP, maybe even MBA or iPad.

 

[alexktz]

Yeah, Mac drivers are garbage in comparison to Windows...

Here is the link to that update, I wonder if 5500m users could install it and update the drivers. I have poor internet connection right now to check it myself: https://support.apple.com/kb/DL2044

Nope, will not install (I have both laptops side by side today - the 5500m goes back tomorrow).

 

[alexktz]

I think you guys are going to like the further testing I've done today. I even borrowed a 38" ultrawide from a friend so that I could test these configurations.

The TLDR is that no matter what configurations I tried I couldn't make the GPU exceed 12w at idle. The exact same configurations with the 5500m model resulted in 20+w as we all know. Two monitors seems to be the sweet spot for energy usage but there's not much in it really.

All tests were conducted using an Anker dongle (linked in config 2) using the Thunderbolt ports on the left hand side of the computer. I did not use the ones on the right at all. Both displays were simultaneously hooked up to the left with no bandwidth issues.

I can only assume whatever driver bug was in the one that is currently shipping for every model except this one is borked and this one has a fix. I'm extremely happy and for now, this is the perfect laptop. Expensive, yes, but perfect - I still miss the SD card slot though

# Config 1

• Single external display over 1 usb-c cable
  • LG U3419W - 3440x1440

Results

• Lid open: Max 12w | Average ~8w | Low 7w
• Lid closed: Max 9w | Average ~6.5w | Low 4w

# Config 2

• Two external displays
  • 1 LG 38UC99-W 3840x1600 (connected via usb-c - full res @60hz)
  • 1 Dell 24 inch 1920x1080 (connected via HDMI using this Anker adapter @60hz)

Results

• Lid open: Max 11w | Average ~8w | Low 7w
• Lid closed: Max 8w | Average ~5.5w | Low 3.8w

# Config 3

• Two external displays
  • 1 LG 38UC99-W 3840x1600 (connected via usb-c - full res @60hz)
  • 1 LG C9 4k OLED TV (connected via HDMI @30hz using Anker dongle linked above)

Results

• Lid open: Max 11w | Average ~8w | Low 7w
• Lid closed: Max 11w | Average ~6w | Low 4w

 

[ludalex]

I’ve been using it both closed/open. Usually open for audio production work for extra screen real estate and closed for general web browsing. It has been getting hotter where I’m located so I've been seeing:

(sudo powermetrics)

MBP Open
CPU Die: ~55-85C
GPU Die: ~55-78C
Fans: ~1800-3000 rpm.

MBP Clamshell (Closed)
CPU Die: ~50-75C
GPU Die: ~53-69C
Fans: ~1800-2500 rpm

This was just the min/max after watching terminal while working on light audio work in Ableton with a Twitch stream and a few other tabs open/running simultaneously. Using a 1440p 165Hz monitor and MBP’s display set to “Default for display”. Obviously nothing scientific, but it’s a general idea of what I’m experiencing. Might be worth mentioning that I’m using stock MacOS Catalina 10.15.5 fan/CPU management (not using apps like Macs Fan Control/Turbo Switcher).

When just web browsing with multiple tabs in open mode, I’m consistently seeing CPU/GPU temps of ~50-55C at 1800rpm. CPU temps are almost always higher than GPU temps. I don’t know if this but is relevant to my stats but I’m running all my peripherals, monitor, and power over a single cable from my HP Thunderbolt 3 G2 dock (It’s the only cable plugged into my MBP).

edit: Capitalization + more context.

How do you connect your TB3 dock to the gaming monitor?

 

[jc_9]

Damn, that's some good news for the 5600M!

12W (or 24% TDP) is now very acceptable for 3 total displays at idle. So is it really worth paying $800 just for having a silent desktop experience? Maybe not, but less wattage means lower temps and better longevity. The impressive GPU future-proofing helps when thinking the added cost, but I'm not sure if future-proofing is the better term considering the ARM macbooks are coming soon...

I would have much prefered AMD releasing a software fix for the 5300M and get the base 16 refurbished, but I need a macbook pro now..!

 

[alexktz]

Damn, that's some good news for the 5600M!

12W (or 24% TDP) is now very acceptable for 3 total displays at idle. So is it really worth paying $800 just for having a silent desktop experience? Maybe not, but less wattage means lower temps and better longevity. The impressive GPU future-proofing helps when thinking the added cost, but I'm not sure if future-proofing is the better term considering the ARM macbooks are coming soon...

I would have much prefered AMD releasing a software fix for the 5300M and get the base 16 refurbished, but I need a macbook pro now..!

I don't see how if they fixed it for the 5600M that those fixes won't trickle down to the other models soon.

 

[jc_9]

I'm thinking the HBM2 vs GDDR6 memory is the major reason for the wattage difference. Hopefully I'm wrong!

 

[mrmachine79]

I don't see how if they fixed it for the 5600M that those fixes won't trickle down to the other models soon.

Because it's a hardware problem. The GPU needs to run at higher speed to sync multiple displays at different resolution and refresh rate. The 5600m has 2x faster memory. Maybe it doesn't need to run the memory at max speed (half speed is fast enough), or maybe like the 2.4ghz i9 it just generates less heat for a given workload. Seems likely given it's still 50w TDP and 75% faster.

Either way, good news for those who can afford it. My eGPU with 5500 xt was still much cheaper and already an unnecessarily expensive "fix".

But the 5600m would have been comparable cost and performance to an eGPU with 5700 xt, without the eGPU quirks.

I bet if you complain long enough and jump through enough hoops including an email to Tim after hitting the end of the line with regular support, they will offer a replacement and you can pay the difference to upgrade. They replaced my 2019 15 inch after 11 months with a new 16". But it took 9 weeks of dealing with support and trying twice to repair it.

Even with the crazy high cost, I wish the 5600m had been available before I chose the upgrade spec and before I purchased eGPU.

 

[alexfc]

I'm thinking the HBM2 vs GDDR6 memory is the major reason for the wattage difference. Hopefully I'm wrong!

I think maybe it has an effect, but not to the extent we’re seeing here. Some people are reporting no problems on wattage with 1 monitor. The huge jump, for them, seems to happen when a 2nd monitor is connected. Also, we’re seeing that on Windows it can run just as good as the HBM2, so I don’t think it’s all about the memory, but more of a software related problem.

Other than that, yeah, 5600 has 2x faster memory. It’s more powerful, but If that’s the logic we’re using here (more powerful, so it performs better and solves heating issues), then why our 5500M is getting worse results than 2016-2018 MacBooks Pro, which have less powerful dGPUs? So, this logic only applies for the 5600M vs 5500M? Also, it’s just one or two monitors. We’re not taking about real high-demanding tasks where the extra power would have an impact.

 

[jc_9]

I think maybe it has an effect, but not to the extent we’re seeing here. Some people are reporting no problems on wattage with 1 monitor. The huge jump, for them, seems to happen when a 2nd monitor is connected. Also, we’re seeing that on Windows it can run just as good as the HBM2, so I don’t think it’s all about the memory, but more of a software related problem.

HBM2 memory is known for having lower power consumption than GDDR6, so it's safe to assume it plays a major part in the numbers alexktz shared us.

"The huge jump" is not specific to the 5300M / 5500M, even nVidia desktop cards do this. For example on my Windows desktop, I have a Geforce GTX 980 running 2 x 1440p@60hz and one 1080p@60hz.

One monitor: 11.5% TDP
Two monitors: 11.7% TDP
Three monitors: 27.3% TDP

It's a desktop GPU thought and the fans stays silent (1400RPM) even when using 27% TDP but I remember not having this huge jump with the same monitors a couple years ago.