Tableau & Apple silicon M1 support in November 2021

0:00Hey it's Tim here, I just wanted to do a

0:01very brief video covering Apple Silicon

0:04support for Tableau.

0:05Essentially a while back I did a video

0:07about this and it kind of worked and I just

0:09wanted to do a

0:09bit of an update because a few people have

0:11asked me on Twitter whether they should buy

0:13some of the

0:13new M1 based Macs to work with Tableau. So

0:16I'm just going to do a very brief video, it

0:18's not

0:18going to take long I'll be pretty honest

0:20with you, it's still not an ideal solution.

0:22Okay so here I

0:23am, you can see that I'm in macOS Monterey,

0:26I've got a brand new MacBook Pro, I've got

0:28one of the

0:29new ones, the M1 Macs based one with 32GB

0:31of RAM, basically a sort of a mid-level

0:34tier spec. I got

0:36it because obviously I edit videos so I

0:38went for a slightly specced up version but

0:40it's got 16GB of

0:41RAM, sorry it's got 32GB of RAM not 16GB

0:45and it's got a 1TB hard drive. So it's very

0:49comparable to

0:49what I had before, essentially before this

0:52I had another MacBook Pro 32GB of RAM, it

0:54had the base

0:55sort of level GPU and the reason I got that

0:57is because I do a lot of different things,

0:59of course

0:59I work in analytics but I also do a lot of

1:01video editing and creative work outside of

1:03that as well.

1:04And so what I will say a little later on is

1:06that I think it's always a good idea to get

1:0832GB of RAM

1:10if you're on a Mac and you're trying to

1:11work in analytics, I'll explain why in a

1:13second. But you

1:14can see my spec here, now you can also see

1:16here that I'm in Tableau 21.3.3, this is

1:19the latest

1:20build of Tableau and so the simplest case

1:23here is that I've just done a very simple

1:26vanilla install

1:27of Tableau here, I've just done the default

1:29install and typically if you did the

1:30default install

1:31one of the things you'd get is a driver for

1:34Microsoft SQL Server and I think another

1:36driver

1:36for Postgres, I'm never sure which two it

1:38is but there's two standard, I think it's

1:39Oracle or

1:40Postgres, one of those two comes as part of

1:42the default installation. Now what's

1:44interesting is

1:45I've actually run this installer multiple

1:47times and if I go ahead and just try and

1:48connect to one

1:49of the most standard databases here you'll

1:52see that it asks me for the driver and I've

1:54actually

1:54run this installer multiple times and this

1:56is actually a bit of a sticking point. Now

1:58there's

1:58a couple workarounds to this, I have been

2:00able to get this to work but it's kind of f

2:02iddly and it's

2:03not supported, Tableau is not officially

2:05supporting this so it's not ideal. If

2:07however you install

2:08Parallels on your M1 base Mac or your Apple

2:11Silicon Mac and then on that version of Par

2:14allels

2:15you run an ARM version, ARM version of

2:17Windows 11 and then you install Tableau on

2:20that. For some

2:22bizarre reason the emulation seems to sort

2:24of work out fine so here I am, I've

2:26switched over,

2:27you can see I've got the exact same version

2:2921.3 64-bit, this time a Windows install of

2:32the same

2:33thing and if I go ahead and I just switch

2:35over to the data connections here you can

2:37see that I'm

2:38actually already connected to the SQL

2:39Server and I've played around with this, I

2:41've done some basic

2:42work and it works and so I'm able to

2:43connect to the data. Now what I'm not

2:45saying here is that

2:46this is supported and it's going to work

2:48fine and you should build stuff off it, all

2:50I'm saying is

2:51that it's still quite early days for this

2:53technology and so this is sort of the work

2:55you

2:55have to do if you want to use one of these

2:57machines and still do your daily work, you

2:59have

3:00to invest some time figuring out what doesn

3:02't work and what doesn't work and that's not

3:04ideal because

3:05you should really just be able to get on

3:06and do your work, you don't want to have to

3:08spend time

3:08sort of just quickly doing some analysis

3:10only to find you've hit a bug. A couple of

3:12other bugs I

3:13found if I go over to this one, if you try

3:15and do activation via Tableau Online or

3:18Tableau Server,

3:19it doesn't quite work on the M1 based

3:20version here but it does work completely

3:22fine on the Windows

3:24based version. Another detail is it also

3:26depends on the version, for example when I

3:28try that license

3:29based activation on a version from 2020 it

3:32doesn't work at all, when I try it on 2021.

3:353 it works

3:36completely fine, I tried it in 2021.1 it

3:39didn't quite work and I just didn't sort of

3:41do any sort

3:41of scientific test to see exactly which

3:43versions work and which ones don't and so

3:45you're probably

3:46wondering well what is supported and what

3:48isn't supported? Well the short answer is

3:50none of it's

3:50supported, if you go over to the Tableau

3:52community forum there's actually a really

3:54good post here,

3:55I'll put this in the description below as

3:57well so you can go directly to it in the

3:59links and there's

3:59an update here from David Brown from Table

4:01au essentially covering that look as of June

4:032021

4:05this hasn't been updated I don't believe

4:07since, it's not officially supported so

4:09there are things

4:10that are going to work, that's absolutely

4:12the case you can run it, install it, it

4:14runs on Rosetta

4:15absolutely fine but you are going to hit

4:17bugs and those bugs are not supported,

4:19Tableau's not going

4:20to go out their way to fix them until they

4:22officially either compile a version that

4:25works

4:25on Amble silicon natively or I assume they

4:28do enough sort of coding and work to this

4:32version

4:32that it runs fine under Rosetta, quite a

4:34few applications have actually done this,

4:36if you look

4:36at some of the Adobe apps those run

4:38absolutely fine under Rosetta because Adobe

4:40have done the

4:41work to make sure that it's going to work

4:43fine under Rosetta but in other cases other

4:45applications

4:45just flat out don't support Apple silicon

4:47so if you're thinking of getting an Apple

4:49silicon Mac

4:50be super careful if you're going to be

4:52using this Mac as a daily driver for your

4:54work and you just

4:55need to get stuff done I'd recommend you

4:57hold on to the laptop you've got you

4:58probably don't want

4:59to buy a new Intel based Mac at all but you

5:02're also probably itching to buy something

5:05new and

5:05maybe hold out for that so just hold out a

5:07little bit more, Tableau's not really said

5:10when they're

5:10committing to update this I hope it's soon

5:12I hope it's by this time next year because

5:14by then the

5:15Apple silicon transition will have been

5:17complete and Apple should only be really

5:19selling Macs with

5:21Apple silicon in them only so I'm hoping

5:23that Tableau's sort of matching that pace

5:25with

5:25development. The other sticking point is

5:28that of course the Mac OS versions also

5:30change every year

5:31so with Big Sur then Catalina and now Mont

5:33erey there's always a few bugs that sort of

5:35sneak in

5:36as Apple changes the base level systems as

5:38well so it's always a sort of cat and mouse

5:40game to get a

5:40stable version of Tableau sometimes and

5:43even if you had an Intel based Mac you'll

5:44still come across

5:45a few bugs in the latest release so bugs

5:47aren't sort of going to always disappear

5:49they're never

5:50going to be perfect but you kind of have to

5:51work within the realms of what you know

5:53works so

5:54if you're prepared to do some testing and

5:56you've got a very narrow use case for this

5:57maybe you're

5:58just doing exploration this is going to be

6:00great. I will also say that this doesn't

6:02stop you from

6:02going on and using WebEdit you can of

6:04course use WebEdit to just connect to

6:06something on server or

6:07Tableau online and just get working so that

6:09's something you should consider. Now the

6:12other

6:12thing I wanted to touch on is if you're

6:13going to buy one of these Apple silicon Mac

6:15s what should

6:16you get well my recommendation is to get a

6:1932 gigabyte machine of RAM now 16 gigabytes

6:23used to

6:23be the staple for most laptops and it still

6:25is in most businesses now the problem you

6:28're getting is

6:28that a lot of these new machines a lot of

6:30new operating systems work on the

6:32assumption that RAM

6:34is now relatively cheap and therefore it

6:36should be in general available and abundant

6:39in machines and

6:40so you're seeing a couple of new trends

6:42happen you're seeing SSDs get faster and

6:44you're seeing

6:45RAM get faster even something like Windows

6:4811 is optimized to use a new type of RAM

6:51called DDR5 and

6:52so you get into this world where RAM and

6:55hard drive space are getting quite

6:57accessible and so

6:58when developers build applications they're

7:00increasing the amount of resources they

7:02need.

7:03So if I give you a simple example let's say

7:05you get a 16 gigabyte machine okay the base

7:07OS

7:08typically takes five to eight gig of RAM so

7:10that's already half of your RAM gone if you

7:13're

7:13sort of using it heavily if you open up

7:15Chrome maybe let's say you've got six tabs

7:17that's maybe

7:18another gig to two gigs gone okay so you

7:20started with let's say eight now you're on

7:2210 gig of RAM

7:23you've only got six left as soon as you

7:25open up Parallels to emulate Windows a lot

7:27of that is

7:28going to go maybe you use up 4 gig leaving

7:30only with 2 gig of RAM left and then if you

7:32've got

7:33anything else open maybe it's Excel maybe

7:35you've got a database running any other

7:37piece of work

7:37going on it's going to use up that RAM

7:39really quickly and what you don't want to

7:41do is get in

7:41this situation where your computer is

7:43having to do something where it's swapping

7:45information from the

7:46memory to the SSD. Now on Apple Silicon

7:49this actually happens incredibly fast it's

7:52it's

7:52actually silly the SSD speeds on Apple

7:54Silicon machine are just incredible so you

7:57might not even

7:58notice that this is happening if you do get

8:0016 gig of RAM but you will get much better

8:03performance

8:03if you have more RAM than you have you know

8:06SSD space being used by swap so my

8:08recommendation is

8:09to get the 30 gig of RAM it sounds a lot it

8:12sounds like you're kind of pushing the

8:14boundaries but if

8:15you can get hold of it that's definitely

8:17the spec to go for in terms of everything

8:19else you can just

8:20get an M1 Pro essentially you don't have to

8:22worry too much about the M1 Max. M1 Max is

8:24really only

8:25for video creators and people doing high

8:27intense creative work where the GPU cores

8:30are really

8:31useful analytics doesn't tend to use a lot

8:33of the GPU cores and the M1 Max have a

8:36neural engine for

8:37some of the machine learning based activity

8:39so if people are writing software that's

8:41optimized for

8:42that that should be using that side of the

8:44chip in the long run so really you only

8:46need an M1 Pro

8:48and you could probably get the base level M

8:501 Pro that will be completely fine but make

8:52sure you get

8:5330 gig of RAM get a decent hard drive 512

8:56is the absolute minimum one terabyte is my

8:59recommendation

8:59again so you can have a lot of applications

9:01on your machine but 512 is probably what

9:03you'll get

9:03for work that's totally fine if you install

9:05like 10 applications that you use every

9:07single day

9:08and you'll have plenty of space left over

9:10to do your day-to-day work okay and that's

9:12pretty much

9:12it so yeah I just wanted to do this simple

9:14update on Apple Silicon M1s just for anyone

9:17who's sort of

9:17thinking of getting into these super

9:19excited about technology and they use Table

9:21au just hold out if

9:22you want to use this as your daily driver

9:24or if you do be sure that you spend a bit

9:26of time

9:26investing in making sure you have a setup

9:28that works the other thing to bear in mind

9:30is if you

9:30get this for work and your computers are

9:33fairly locked down every time you come

9:35across a bug

9:35you're going to have to be on the phone to

9:37it support and they're not going to like

9:38that so

9:38they might not even let you get this

9:40machine in the first place so just a few

9:42things to think

9:43about and that's pretty much it for me

9:44thanks for watching and I'll catch you in

9:46the next video let

9:46me know what else you'd like to see on an M

9:491 chip from an analytics perspective I've

9:51been able to

9:52get SQL Server to work in Parallels I've

9:54installed a few other things like Docker as

9:56well so if

9:56you've got an idea something you'd like to

9:58run on Apple Silicon that you use with

9:59Tableau let me know

10:00in the comments and I'll go ahead install

10:02it and see what works and I'll try and

10:04reply to your

10:04comment as soon as I can. Thanks for

10:06watching see you in the next one.