Apple M1 + QuPath

Continuing the discussion from Qupath on Apple silicon ARM M1:

Hey Alan, Hey Pete,

@Pete: if you allow me, I would like to append to this topic a short general question: Is it preferable to use a new Mac with the M1 technology (e.g. when compared to Intel I7), or will this likely cause trouble on the long hand?

@Alan: How is your experience with the handling? Would you recommend to stick with the classic ones, or did you experience major improvements with the new chip?

Thanks for the advice!
Thiemo

I really don’t know. There are some interesting opinions here:

I currently use a 2013 iMac most of the time, and a 2014 MacBook Pro sometimes (along with Windows & Ubuntu rarely). I plan to get a new computer this year, when new models are available – probably a MacBook Pro, although possibly an iMac.

Either way, I personally intend to go for Apple Silicon to (I hope) maximize the lifespan. If QuPath problems emerge, I’ll try to resolve them then. I expect Java support will improve with Java 17, due in September (based on this).

1 Like

Hi @TMM-98,
I haven’t had any issues since increasing the RAM allowance in the Qupath settings to 8GB. Compared to my old 2014 intel macbook it maybe a little faster carrying out Qupath tasks but the big benefit seems to be in running parallel tasks. For example, I can run a project analysis in Qupath and also use multiple other applications like web browsers, email client, office apps etc - previously I would close all other apps and run project analysis overnight. Performance is good and presumably as and when new software optimised for the chip comes out it will be even better. The caveats here are that I am not comparing a new intel model with a new M1 model and also, I use a subset of Qupath features. Happy to check if a particular feature works if that is of any help? :slight_smile:

2 Likes

my 2015 mac book air is still kicking and it use it most of my coding. but i really want to check out how efficient rosetta is in the new M1 macs


Check out some of my projects and tutorials in EduHelphub

1 Like

Specifically I was wondering whether neuronal network-based tasks are being carried out with the propagated crazy speeds. It drives me crazy that training and application of pixel classification takes forever at the moment and effectively shuts down my Mac completely.

Though I have now found out that it is only possible to connect one external display at a time, which is a clear disadvantage for me will probably prevent me from upgrading any time soon.

But thanks for the numerous insights!