Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
New build out!

1.5.00.12215 (osx-x64 + osx-arm64) - published on Monday, May 2, 2022 at 10:07 PM with 203 MB: https://statics.teams.cdn.office.net/production-osx/1.5.00.12215/Teams_osx.pkg

Memory usage during a call with in-built webcam on.

1651583024293.png


EDIT: Webcam not working and screen sharing isn't either.

12110 is working pretty great here, although I don't really share my own screen much. Thanks for posting the most recent builds here. One thing I've noticed, if you try to use the 3 dots menu and check for updates, it updates you to the Intel version. Have to make sure now to check for updates within Teams itself.
Yeah since these are the insider builds, we are better off just manually upgrading until they release a Preview client.
 
Last edited:
My teams auto launches from the browser [chrome]; since doing this testing I'd like to stop it doing this so I can choose if I want to be in the browser or not (this is how it used to be before I told it not to ask me)
Does anyone know how to set it back to the 'ask' setting? My google-fu has let me down
 
Anyone know what the difference is between these experimental builds, and "continuous deployment" versions at the top? Seems like the continuous deployment version numbers are newer, although there's no M1 build of those.
 
Anyone know what the difference is between these experimental builds, and "continuous deployment" versions at the top? Seems like the continuous deployment version numbers are newer, although there's no M1 build of those.
I've been using these builds for a few years but exploration (-E) builds seems to be based on newer versions of Electron. At the moment it's on Electron 16 so there's now an M1 build for that since April 2022.

Continuous Deployment (CD) seems to be automated daily (Canary) builds of their master branch whenever a Microsoft employee makes changes to the source code on Azure DevOps. At the moment it's on Electron 10 and these builds looks to be the same builds (at least a few of the release candidates are) that are pushed to the public for everyone.

Both versions look to be internal builds meant for Microsoft employees and maybe those in the Technology Adoption Program.
 
Last edited:
  • Like
Reactions: Zorn
Not sure if I can get these directly or search the ms for this and future builds, but thanks so much for posting!
 
Just curious @MrGunny94, how much CPU/GPU extra usage does screen sharing/webcam add on the versions you managed to get it working?

Cause for me, still at production version though, as soon as I start screen-sharing and/or specially webcam temps start to rise due to higher CPU/GPU usage, not dramatic, but within a certain amount of time, they rise up by 10ºC after 10-15 mins. I was wondering if the native Apple Silicon version is improving that.
 
  • Like
Reactions: MrGunny94
Just curious @MrGunny94, how much CPU/GPU extra usage does screen sharing/webcam add on the versions you managed to get it working?

Cause for me, still at production version though, as soon as I start screen-sharing and/or specially webcam temps start to rise due to higher CPU/GPU usage, not dramatic, but within a certain amount of time, they rise up by 10ºC after 10-15 mins. I was wondering if the native Apple Silicon version is improving that.
CPU/Memory usage is much better but there’s still an increase in GPU.

However both webcam and screen sharing are broken atm and I’m using web app for calls.

In the web app I still have the same experience where 2 cores are fully user and temps go all the way up to 62c
 
  • Like
Reactions: SirKeldon
CPU/Memory usage is much better but there’s still an increase in GPU.

However both webcam and screen sharing are broken atm and I’m using web app for calls.

In the web app I still have the same experience where 2 cores are fully user and temps go all the way up to 62c

For me webcam isn't broken, but people tell me I have almost a black & white effect. For as dramatically better as this performs, they can deal with seeing my picture slightly off color ha.
 
  • Like
Reactions: Macintosh IIcx
For me webcam isn't broken, but people tell me I have almost a black & white effect. For as dramatically better as this performs, they can deal with seeing my picture slightly off color ha.
Same.

They saw me B&W but strangely my lips were slightly saturated. As if I wore lipstick.

I would let them deal with me being B&W, but I can't deal with myself wearing lipstick.

Downgrading.
 
I remember when AS came out many were blithely insisting that the transiton to AS would be seamless, because Intel apps could be recompiled to run on AS "at the touch of a button". I tried to tell them it wouldn't work that way for anything reasonably complicated but, as is often the case, that fell on deaf ears.
 
  • Like
Reactions: Macintosh IIcx
I remember when AS came out many were blithely insisting that the transiton to AS would be seamless, because Intel apps could be recompiled to run on AS "at the touch of a button". I tried to tell them it wouldn't work that way for anything reasonably complicated but, as is often the case, that fell on deaf ears.
Well MacOS Intel Apps can be recompiled to run on AS "at the touch of a button" assuming of course you have the source code to your Application and any third party dependencies and you are using the native Apple tools to build it (e.g. a Swift App using the Swift Package Manager for its dependencies).

The core Office 365 applications are all already native to Apple Silicon, Teams is an outlier. Since it an Electron app, most of the code doesn't even need to be recompiled because it is Javascript. Microsoft needs to package a version of Electron runtime for Apple Silicon along with any binary dependencies.

Teams is not the only Enterprise focused application to lack native support for AS, I suspect the probable is more organizational than technical.
 
Well MacOS Intel Apps can be recompiled to run on AS "at the touch of a button" assuming of course you have the source code to your Application and any third party dependencies and you are using the native Apple tools to build it (e.g. a Swift App using the Swift Package Manager for its dependencies).

The core Office 365 applications are all already native to Apple Silicon, Teams is an outlier. Since it an Electron app, most of the code doesn't even need to be recompiled because it is Javascript. Microsoft needs to package a version of Electron runtime for Apple Silicon along with any binary dependencies.

Teams is not the only Enterprise focused application to lack native support for AS, I suspect the probable is more organizational than technical.
All the evidence I've seen indicates that "at the touch of a button" significantly oversimplifies the technical challenges, particularly for large (say > 1 GB installed), complex applications:

One developer of a medium-sized (91 MB installed) popular Mac app (who I won't quote directly here, since this was said on a non-public forum) explained that it was easy to get it building once you'd compiled the libraries, but that the latter presented some challenges. So that's clearly not "at the touch of a button". He also mentioned there's a difference between getting it building and getting it actually working, because of other changes Apple may make.

In addition, reports in the tech media (I'll try to find the link, don't have it handy) indicated that Apple (wisely) sent engineers to work with MS and Adobe well before the DTK was released, so that Office and CC, which are both large (Word, Excel, PP, and Outlook are 1-2 GB each, installed), complex, and commercially important, would be ready when the first AS machines were publicly released. I.e., they recognized that the lead time between the release of the DTK and the public release of AS would be insufficient to create working native AS versions of these. Logically, if AS versions of Office and CC could be created "at the touch of a button", this would not have been necessary.

Finally, consider Wolfram Mathematica, which is a large (11 GB installed), complex program for doing symbolic and numerical calculations. Mathematica is by far Wolfram's most important app, and Mac users are a key part of their customer base, so they made a big push to get a native AS version of Mathematica working as soon as possible. Yet they weren't able to deliver one until July 2021, which is 13 months after the DTK was released. And even now, a major release later, it still doesn't run as well on AS and it does on Intel. This is technical. It's not organizational.
 
All the evidence I've seen indicates that "at the touch of a button" significantly oversimplifies the technical challenges, particularly for large (say > 1 GB installed), complex applications:

One developer of a medium-sized (91 MB installed) popular Mac app (who I won't quote directly here, since this was said on a non-public forum) explained that it was easy to get it building once you'd compiled the libraries, but that the latter presented some challenges. So that's clearly not "at the touch of a button". He also mentioned there's a difference between getting it building and getting it actually working, because of other changes Apple may make.

In addition, reports in the tech media (I'll try to find the link, don't have it handy) indicated that Apple (wisely) sent engineers to work with MS and Adobe well before the DTK was released, so that Office and CC, which are both large (Word, Excel, PP, and Outlook are 1-2 GB each, installed), complex, and commercially important, would be ready when the first AS machines were publicly released. I.e., they recognized that the lead time between the release of the DTK and the public release of AS would be insufficient to create working native AS versions of these. Logically, if AS versions of Office and CC could be created "at the touch of a button", this would not have been necessary.

Finally, consider Wolfram Mathematica, which is a large (11 GB installed), complex program for doing symbolic and numerical calculations. Mathematica is by far Wolfram's most important app, and Mac users are a key part of their customer base, so they made a big push to get a native AS version of Mathematica working as soon as possible. Yet they weren't able to deliver one until July 2021, which is 13 months after the DTK was released. And even now, a major release later, it still doesn't run as well on AS and it does on Intel. This is technical. It's not organizational.

The size of an app is not proportional to the complexity of porting it between x64 and ARM64. MS Office is large and complex and written mostly in C++ which does facilitate the writing of non-portable code. However, the Office team ported the Windows version of the apps from x64 Windows to ARM64 Windows over a decade ago. They also maintained 32bit and 64bit Windows versions of the apps for years and iOS versions of the apps. I doubt they needed much help from Apple to support ARM64 Macs. Unlike PowerPC or the 68000 CPUs in the older Mac, the ARM64 in the new Macs are little endian like the x64 and x86 architectures.

Wolfram Mathematica was probably optimized for the Intel instruction set and probably uses Intel SIMD instructions which are of course not portable.
 
The size of an app is not proportional to the complexity of porting it between x64 and ARM64. MS Office is large and complex and written mostly in C++ which does facilitate the writing of non-portable code. However, the Office team ported the Windows version of the apps from x64 Windows to ARM64 Windows over a decade ago. They also maintained 32bit and 64bit Windows versions of the apps for years and iOS versions of the apps. I doubt they needed much help from Apple to support ARM64 Macs. Unlike PowerPC or the 68000 CPUs in the older Mac, the ARM64 in the new Macs are little endian like the x64 and x86 architectures.

Wolfram Mathematica was probably optimized for the Intel instruction set and probably uses Intel SIMD instructions which are of course not portable.
I think your technical arguments are obscuring what is the essential bottom line here: that, when people used phrases like "at the touch of a button", they were incorrectly trivializing how technically challenging it can be to port complex apps to AS. In addition to Mathematica, I suppose AAA games would be another example. I don't understand why you're not simply willing to acknowledge this.
 
I think your technical arguments are obscuring what is the essential bottom line here: that, when people used phrases like "at the touch of a button", they were incorrectly trivializing how technically challenging it can be to port complex apps to AS. In addition to Mathematica, I suppose AAA games would be another example. I don't understand why you're not simply willing to acknowledge this.
I think if you are going to argue that something is technically challenging, then the technical arguments are what is relevant. I have ported C and C++ code between multiple compilers, operating systems and CPU architectures. It doesn't get much easier than MacOS x64 to MacOS ARM64. The OS is the same, the compiler is the same and the byte order is the same.

Some applications will certainly have challenges, particularly those with complex third party dependencies or those specifically optimized for x64, and any assembly code will have to be re-written. AAA games are challenging because most of them haven't even been ported to MacOS or Metal to begin with. Many of them target DirectX.
 
  • Like
Reactions: michalm
I think if you are going to argue that something is technically challenging, then the technical arguments are what is relevant.
Not when the technical arguments are being used to obscure instead of illuminate.
Some applications will certainly have challenges, particularly those with complex third party dependencies or those specifically optimized for x64, and any assembly code will have to be re-written.
So here you're finally willing to acknowledge what I was saying, yet getting you to admit to this was like pulling teeth.

With all due respect, here's the problem I have with what you've been doing. In your initial response to my post, if you wanted to act in a collegial, helpful manner, you could have instead written something to this effect right from the start (maybe it's not exactly correct, but you get the idea):

"If you have a self-contained set of code designed for MacOS under Intel, you actually can recompile it to run under AS 'at the touch of a button'. Having said that, the essence of what you are saying is correct, since many apps aren't self-contained sets of code. Instead, they are linked to libraries, or have other dependencies, that must also be ported to AS, and this can create challenges. Further, sometimes they are specifically optimized for x64, which means assembly code must be rewritten. So, yes, it is a misleading trivialization of the process to say that apps can be ported 'at the touch of a button'."

Let me give an example from my own field. I'm trained in physical chemistry, and frequently participate on chemistry forums. Suppose someone asks the following: "Isn't the only reason we care about ΔG_system is that it act as a surrogate for ΔS_universe, and it's the latter that really matters when it comes to determining spontaneity"?

Here's the helpful and collegial response: "Yes, that's essentially correct. However, you should note that ΔG_system = –TΔS_universe holds only under conditions of constant T and p (which are, granted, the most-commonly specified conditions). Under other conditions, you would need to use a different state function as a surrogate for ΔS_universe (e.g., at constant T and V you would use ΔA_system). Since you are working at constant T and p, ΔG is applicable."

And here's the non-helpful game-playing response that, while techncially correct, obscures the essence of what's going on: "It's not generally the case that ΔG_system can be used as a surrogate for ΔS_universe. For instance, if you are at constant T and V, you can't use ΔG_system as a surrogate for ΔS_universe."

The person who gives the latter response generally isn't interested in being helpful and promoting understanding. Rather, they're more interested in telling the other person they're wrong. I'm sure, being in a technical field yourself, you've encountered that type of individual.

Now ask yourself which of the two of these your original response most closely mirrored.
 
Last edited:
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.