Hacker News Reader

Comment by miohtama | Hacker News Reader

Sounds like this explains why Docker eats thru Mac battery so fast

Checkout App Tamer[1] to control the priority of processes as well as core assignment on the M series chips:

> App Tamer can take special advantage of Apple Silicon powered Macs, which have two different types of processor cores. Use it to automatically run busy background apps on the M1 or M2's efficiency cores to save power, leaving the performance cores for the apps you want to run fastest.

1. https://www.stclairsoft.com/AppTamer/

I get 6h or so running quite a few workloads in Kubernetes on Docker For Mac on Apple Silicon. Intel was < 2h. The rest of my team has similar experiences. Are you sure it's Docker causing your battery issues?

Docker is definitely a problem, although newer versions have implemented new functionality that may have significantly reduced this. I can't really tell however, as I'm stuck on too many Zoom calls all day long and Zoom absolutely murders my battery life.

I haven’t tested it personally but you may be able to get some of that battery life back by using the web version of Zoom in Safari. That tends to force video chat services to use hardware accelerated video codecs which are considerably easier on battery.

And that's not just for Mac's, the same can help on Linux devices. It's somewhat hilarious that running a non-native app gets you better hardware accel but here we are.

Maybe the web browser is the OS now. Does that make the OS just a hardware abstraction layer?

I think that's generally accepted as true for some classification of people.

That's why chromebooks are an idea that was perhaps too early (and hardware too anaemic)

I don't know about zoom specifically, but aren't most web conferencing "native" apps just electron anyway?

Indeed they tend to be however the underlying electron base tends, imo experience, to be outdated and comparatively hard to configure

Try orbstack

Does it make use of the E cores on Apple Silicon?

Since all docker implementations use a VM on Mac, and according to this article VMs only use P-Cores, I have to assume it only uses the P cores.

This article is based on an M1 Max but I think M2 Max behaves differently here since it has more E-cores. Don't quote me though.

(dev here) Yes. I've spent quite a bit of time on scheduling. It's more nuanced than what this article says and there are ways to influence it (but sorry, can't share too many details). OrbStack tends to use E cores more than other virtualization-based apps do on Apple Silicon, especially for heavy workloads.

Thank you! This is really interesting, will give OrbStack a test drive.

Orbstack still seems to rely on docker's virtualization. We need something better than docker to improve on this.

Orbstack runs Docker inside its own lightweight VM. It doesn't use any of Docker Desktop's virtualization. I'm not sure if it handles scheduling any better though.

Reply on news.ycombinator.com