[Feature Request] Native Windows support

[raman1383]

Review Mojo's priorities

• [X] I have read the roadmap and priorities and I believe this request falls within the priorities.

What is your request?

native support for windows.

when will it be available?

What is your motivation for this change?

I use Windows

Any other details?

No response

[melodyogonna]

I could not understand why Modular didn't start the project with Windows.

Because they're not targeting desktop market. Max Engine aims to simplify AI inference and deployment. AI inference is done with GPUs (or other accelerators), coordinated by Linux. Nobody deploys Windows.

Developing on Windows though is another matter, most developers still use Windows. since developers and developer experience is important for Modular you can be sure Windows is important to them too.

[xmaxrayx]

The issue is most local AI on individual hardware are limited to low vram not on speed most AI script that written on python are warped to Cuda so speed not that big issue.

so understandable MOJO will care about people who actually use their product aka Linux server users, mac is based on Unix and doesn't have that variaty on hardware, so it's not hard.

Windows in other hand defferent system, defferent hardware and let's not forget the new thing Arm windows and npu and good luck with snapdragon documentation (as I heard) , so it's not that easy we will be lucky if they did it in 1 year, let's not forget not all windows user devs or care about AI.

1 second saved in individual user machine not the same as saving 1 second on server machine.

On Tue, Jul 16, 2024, 2:03 PM carlos bahia @.***> wrote:

I see. By the business point of view, you have to look what is more lucrative....

On Tue, Jul 16, 2024, 08:59 xMaxrayx @.***> wrote:

I believe because Macos is based on Unix so its easier than windows.

On Tue, Jul 16, 2024, 2:53 PM carlos bahia @.***> wrote:

Hi

By the statcounter (link below), we have:

1. Desktop Operating System Market Share in United States Of America - June 2024

Windows has 62.4% OS X 24.7%

Where Windows is 2.5x bigger than OS X

https://gs.statcounter.com/os-market-share/desktop/united-states-of-america

2. the Desktop Operating System Market Share Worldwide - June 2024

Windows has 72.9% OS X 14.9%

Where Windows is 4.9x bigger than OS X

https://gs.statcounter.com/os-market-share/desktop/worldwide

I could not understand why Modular didn't start the project with Windows.

On Mon, Jul 15, 2024, 13:01 Walter Erquinigo @.***> wrote:

Indeed, our priorities right now are GPU support and satisfying the needs of the market. We know that Windows is important and it's included in the roadmap. As you may imagine, there's a good portion of the market using Windows, especially not in the US.

However, at this point I can't share details about timelines.

— Reply to this email directly, view it on GitHub < https://github.com/modularml/mojo/issues/620#issuecomment-2228855477>,

or unsubscribe <

https://github.com/notifications/unsubscribe-auth/ACNUYO3LIQQ4JXRET3RGBCDZMPW4PAVCNFSM6AAAAAA4P7Z3IWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMRYHA2TKNBXG4>

. You are receiving this because you commented.Message ID: @.***>

— Reply to this email directly, view it on GitHub https://github.com/modularml/mojo/issues/620#issuecomment-2230700469,

or unsubscribe <

https://github.com/notifications/unsubscribe-auth/AYGT5YWU432A3SZBQZUAPODZMUCSBAVCNFSM6AAAAAA4P7Z3IWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMZQG4YDANBWHE>

. You are receiving this because you commented.Message ID: @.***>

— Reply to this email directly, view it on GitHub https://github.com/modularml/mojo/issues/620#issuecomment-2230711067, or unsubscribe < https://github.com/notifications/unsubscribe-auth/ACNUYO36IH3VYJQMTYXMRSTZMUDJDAVCNFSM6AAAAAA4P7Z3IWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMZQG4YTCMBWG4>

. You are receiving this because you commented.Message ID: @.***>

— Reply to this email directly, view it on GitHub https://github.com/modularml/mojo/issues/620#issuecomment-2230719149, or unsubscribe https://github.com/notifications/unsubscribe-auth/AYGT5YXU74JFMRARAUBMWHTZMUD2LAVCNFSM6AAAAAA4P7Z3IWVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDEMZQG4YTSMJUHE . You are receiving this because you commented.Message ID: @.***>

[droy-proxsys]

Just keeping this train chugging along... So, like many poor souls, I work in a company that chains me to Windows (ugh, I know, right?). I really want to dive into this language more, and while I'm not launching major AI models into the stratosphere, we do have some heavy ML use cases. Enter Mojo – potentially a game changer!

Yeah, I get it – Windows is the bane of my existence and probably yours too. But hey, for all of us trapped in this Windows wonderland, can we get a ballpark figure on when Mojo might grace us with its presence? We'd really appreciate it, or at least I would, because I’m dying a little inside every day waiting for this to be openly available.

[inikishev]

WSL works fine, I just wish I could run WSL mojo from normal VSCode instead of having to open separate VSCode in WSL and navigate to my windows files through the scary looking linux filesystem, but that is VSCode related problem

[Jabadahut50]

I believe another turn around solution is support mojo build with --target-triple x86_64-w64-win32 so we can run the execution file in windows OS Currently facing issue when build with this target-triple "mojo: error: Couldn't find msvc toolchain."

Honestly this would be the best stop gap. Being able to at least compile exe's for windows is enough for my purposes and I'm sure for a lot of others as well. Not that native windows support shouldn't come down the road, but at a bare minimum getting exe compilation would allow open source users to start publishing cool projects for windows.

[walter-erquinigo]

There's been a lot of activity here, and I can empathize with most of the opinions in this thread, however, I'll be closing this issue for the time being because we don't have Windows in the short term roadmap. We are still committed to bringing Mojo and MAX to Windows when it's the right moment for the company. That day we can reopen this issue and have a more concrete discussion.

[GameCodeMode]

any specific month so that i can prepare for it?

[Moosems]

@GameCodeMode "we don't have Windows in the short term roadmap"

[Neltherion]

"Second Hand Citizen"🫠

[stimpy77]

There's been a lot of activity here, and I can empathize with most of the opinions in this thread, however, I'll be closing this issue for the time being because we don't have Windows in the short term roadmap. We are still committed to bringing Mojo and MAX to Windows when it's the right moment for the company. That day we can reopen this issue and have a more concrete discussion.

I, too, was monitoring Mojo for Windows support. As a .NET developer, Mojo was one of the ways I was looking to move off .NET taking advantage of the Python ecosystem, that was already a big ask even if Mojo did support Windows.

However, I completely understand and agree with the decision made here. Make Mojo awesome, guys, keep focusing on that. Once it's stable and new development can slow down, we can look forward to engineering efforts coming back to getting it working on Windows (currently about 59% of all Internet developers). If we are in a pinch and need mojo we can of course ~try using wsl or~ work directly in an Ubuntu VM.