I will probably be using a M1 Macbook for my work, which heavily consist of using Matlab+Dynare. I’m very concerned about how my workflow is going to be affected by using a different chip architecture, thus I would be very glad if some of you that have been working with Dynare running on M1 Macbook, could provide me some insight on what has been your overall experience using Dynare and Matlab in such processors. Is it stable? Is performance reduction an issue?
I use Dynare 4.6.4.
My understanding is that with Rosetta 2, everything works smoothly as before. There are still some kinks with the native ARM, but as Matlab still does not have an ARM-compatible version, that should not be much of an issue right now. The related ticket is Compile from source on Mac M1 Apple Silicon architecture (#1780) · Issues · Dynare / dynare · GitLab
@wmutschl may be able to say more.
I’ve been using a M1 MacBook Air since April and with Rosetta 2 both Dynare and Matlab run fine. In the beginning there was an occasional freeze in MATLAB, but that has been resolved with the latest macOS and MATLAB versions. Also there is no performance reduction, but rather a performance increase compared to my “old” Dell workstation Precision 7520. So overall, you should be more than fine with Dynare version 5.0 (but also for older ones by compiling them from source).
@wmutschl That’s good to hear. I have never installed Dynare by compiling the source code locally in my machine, rather used the usual installer. Do I have to compile the code locally for better results? In that case, may you have a guide or something to do so? Thank you!
There is no need for that, there is no computational gain (except for flexibility and code development work) in compiling from source. So you should just stick to the installers we provide on our website. The new release candidate for v5.0 is out just today, so feel free to give it a go. The actual release should (if all goes to plan) be on Friday. The Release Candidate can be downloaded from Index of /beta. I just installed it on my M1 MacBook Air and all works fine.