M1 Power Usage is astonishing!

[Fomalhaut]

Just looking at the output from /usr/bin/powermetrics on my M1 Mini - just incredible!

It's quite often using less than 1W doing normal productivity tasks. I'm running 2 x 1080p screens via DisplayLink, so that might explain to tiny GPU usage, but it has very little impact on the CPU as well.

My Dell T3600 6-core Xeon workstation (with GTX1060 GPU) measures about 200W (from the socket) in normal use. Obviously, the Mini is using more than 1W for the rest of the logic board, and power-supply inefficiencies, but even so, it is an astonishing feat.

Reminds me of the story that the first prototype ARM chips back in the 80's ran even when they forgot to connect the CPU to power, and it was running on current leakage from the circuit board.

 

[dmccloud]

ARM as a platform has always had a major advantage over x86 architecture in terms of battery life. I have been surprised at just how well this machine sips power even when running more intensive applications. Right now, I'm only using 149mW with Firefox, Apple News, App Store, System Preferences, and Activity Monitor running.

 

[leman]

I think it's very interesting to look at full system power consumption figures measured by notebookcheck in their in-depth review. Looking at average and peak power consumption:

M1 MacBook Air 25/30 watts
Tiger Lake Dell XPS 13 40/47 watts
AMD 4800U Yoga Slim 48/58 watts

The AMD CPU specifically with it's double the cores can be up to 40% faster than the M1 in multi-core workloads, but it also consumes twice as much power — quite an interesting fact given that these CPUs are sold as 15W/25W parts. What I find even more remarkable is that the tested Intel and AMD laptops only had a FullHD display panel which consumes less power than a HiDPI one.

 

[dmccloud]

There is another benefit to how efficiently the M1 sips power that I just discovered. I decided to see what would happen when using the 30W charger my MBA shipped with instead of the 61W charger the Pro shipped with. While it does charge slightly slower, it does still charge while using the laptop. Normally if you used a charger with half the wattage as the charger your system came with, it would barely have enough to run the system at idle, and would only prolong the battery life because the system would be drawing more power than the adapter could provide. With this laptop, I'm on the forums while streaming from Spotify, and the battery is charging despite the underpowered charger that I'm currently using.

 

[dmccloud]

This is my current power usage while running Safari (downloading XCode) and Firefox. The machine has been off the charger for around 4 hours, and I still have 91% battery remaining:

 

[AndrewNielsen]

In case anyone wonders in the future... I use my M1 13" MBP with the screen about half brightness for very light work.
• A power meter between the mains power and the charger said that it was using about 6.5 W.
• It lasted a bit over 12 hours until the battery went flat. 4.6 W if the battery is 58.2 Wh.
• Also, I used a Cygnet 27000 MAh, 99.9 Wh battery to power it. It needed about 90 Wh to charge it, measured as above. (I assume that the battery did not fully discharge because it still needed to produce 20V to charge the laptop.) It powered the laptop for 9 hours. Therefore, from the point of view of charging the battery, the MBP "used" 10 W.

The power meter was a Hypertec Power-Mate Lite. In Australia.