DSS on Mac with M1 chip

tgb417

All,

With Apples WWDC 2022 coming up, it has been two years since the announcement of the Apple M1 chips based on ARM. Typical tools like Postgrsql, Anaconda even Java have ARM native versions now. Where are we with installing DSS on an Apple Silicon based Macintosh?

I’m asking because I have another laptop decision to make. Last time I ended up with an Intel based Laptop which I really don’t like to be compatible with the company IT policies. To use DSS in a prototype mode I maxed out the laptops ram and installed WSL2. This has worked fairly well for non production use of DSS, even though from Dataiku’s perspective, I’m coloring a bit outside of the lines.

In my new role I’ll have considerable more latitude to chose the hardware I work with as a laptop, can I install DSS on Apple silicon at this point?

---

Operating system used: macOS Monterey

CoreyS

Please note this updated solution for installing Dataiku on a Mac M1: https://community.dataiku.com/t5/Setup-Configuration/DSS-installed-on-Mac-M1/m-p/27513/highlight/tru...

Sergey

Hi @tgb417

The official statement hasn't been changed: We do not yet support running DSS on the M1 chip.

The limitation is mostly because python itself doesn't officially support M1. Apple claimed that starting with Python3.9 they will patch it to natively work on M1 chip but as you know, we do not yet support Python3.9 as the built-in python for DSS.

The workaround remains the same: Rosetta emulator + Python3.7 as marked here:

https://community.dataiku.com/t5/Setup-Configuration/DSS-installed-on-Mac-M1/m-p/17429/highlight/true#M1652

tgb417

@sergeyd
,

Thanks. and understood. I recognize that if I do this, I'll be coloring a bit outside the lines. That said I've been coloring outside the lines for the past year running DSS in WSL2 and the experience has been OK.

For anyone out there running DSS on Apple Silicon. How bad are things? For example do Plugins work? Is there performance horrible in any area. Do the Basic Graphics and Statistical tools work? Visual Machine Learning and the like?

Baseltov

@sergeyd

When will DSS support 3.9 version of Python? Apple silicon has been released for over 2 years, and are extremely powerful machines that a lot of developers are starting to use again.

I really hope DSS will soon be compatible, as it is taking its time. Almost all other programs support apple silicon natively.

Sergey

@Baseltov

DSS supports py38, py39 and py310 (experimental support) managed code (not built-in) envs starting from DSS10.0.4.

tgb417

@sergeyd
,

Now that we are starting to see python and other support for M1. Are we able to re-visit the DSS on the Apple Silicon question?

https://www.anaconda.com/blog/new-release-anaconda-distribution-now-supporting-m1

--Tom

Turribeach

Tom the thread linked by Sergey shows how to install it on Apple Silicon. I have followed it and it works fine. So far everything I tested has worked great on my M1 Max but I haven't done much to be honest. Like Sergey says I think the main issue is that Dataiku do not yet support Python3.9 as the built-in Python for DSS. So that's probably where Dataiku should focus their attention, not only to support running Dataiku on Apple Silicon but to remove obsolete software (aka Python 2.x) from Dataiku. I doubt anyone installing Dataiku on Apple Silicon hardware will use it on a production / commercial basis, so I think the work around suggested by Sergey's link is good enough for now. But I do see the value on supporting Apple Silicon since Macs are quickly regaining the Professional market segment.