Julia/VSCode on new MS Copilot Surface 7 laptops

I’ve been trying to use Julia/Jupyter in VSCode on the new Microsoft Copilot Plus PC (15" Laptop). Without going into too much detail, if I open VSCode on a Julia notebook, it responds that it can’t find the language kernel. If I uninstall Julia (thoroughly, erasing all files) and reboot, I get it to work, but if I restart the machine it fails again. This is for code that works fine on my Laptop 5. So, it seems that Microsoft has some fixing to do before we can use Julia/Jupyter/vVSCode productively on the new machines.

I’ld love to hear back from others that may have solved these problems. Thanks.

Do you use X86 version of vscode (I assume that it is required since there is no ARM version of Julia for windows) ?

Perhaps I’m wrong, but I assumed I was getting an ARM64 version of Julia from the Microsoft Store. Whatever I got worked for a while.

Could you try PetrKryslUCSD/VSCode_Julia_portable: Portable Julia running in VSCode (github.com)?

Actually, if it is the Snapdragon Elite, it is an ARM architecture. So, if X86 runs, it is probably emulated.

Is it the jupyter extension issue? Have you tried using just jupyterlab? I have an arm version of jupyterlab-desktop at Releases · xgdgsc/jupyterlab-desktop (github.com)

Thanks to everyone who responded. Every time I try to avoid Python, some horrible gravitational force drags me back to it. I much prefer Julia, but the local groups are all Python, the Python ecosystem is huge and active, and Python runs on my beautiful new Surface Laptop 7. Ugh!

If you could escape the language I guess it wouldn’t be called Python?

12 Likes

Good one!

I am looking to buy the new snapdragon laptop as well and this is exactly what I am afraid of. However, in your situation, are you sure you’ve downloaded the ARM versions of all the relevant software? Take a look at your task manager to confirm you are running ARM versions.

You could also work in WSL2 rather than windows – It would install ARM ubuntu natively and you could use juliaup to install a native Julia version as well.

1 Like

Thanks for your comment. I downloaded Julia from the Microsoft Store, in the (vain) hope that they would automatically give me the correct version for my machine. Oops. In addition, it turns out that an official Julia version for Windows/ARM is not planned. Further, I don’t want to mess around with Linux: I have enough to learn trying to simulate fluids.

You’re right: There probably is some way to make it work. But, in my case, the game wouldn’t be worth the candle.

Thanks again.

Well, it would be good if someone who has such a laptop could investigate how to run Julia in the X86 emulation mode. And if there are any issues, help to debug them. First without VSCode, that would be the second level.

1 Like

I’ ve used arm windows emulated x64 version of Julia for nearly 2 years just fine. So I guess the issue could be the vscode juyter extension. And I asked the poster to try just jupyterlab and with no response.

I didn’t mean to ignore your suggestion, which I’m sure was a good one. I’m just very committed to the VSCode/Jupyter toolchain, so I didn’t want to add another tool.

No problem, but for the community it is important to debug the issue in a systematic fashion:

  1. does Julia (command line version installed using Juliaup) work? As far as I understand, yes.
  2. can you install the X86 version of VSCode? If yes, how?
  3. does the Julia plugin of VSCode work?
  4. does the Jupyter extension work? If not, what is the error message?

And if 1…3 work, perhaps change the title of the thread.

1 Like

Julia appears to be working on my Surface Laptop 7 now. I kept asking CoPilot if Julia was available for Windows ARM64. In most cases I got a negative response. But one response told me to run “winget install julia -s msstore” in a terminal and then follow with “juliaup add release”. After that, I’ve been able to run my old Julia notebooks in VSCode/Jupyter on the Surface 7.

The question is: is it running in an emulator, or is it actually an ARM version.

Thanks for your comment. I agree that’s an important issue. However, for the last couple of months my beautiful new Laptop 7 has served only as a paperweight insofar as Julia is concerned. I can use it now - emulator or not - until a proper Windows ARM64 version of Julia is released.

1 Like

Is there a way to target the GPU on those devices? OpenCL.jl on Linux should work, right?

George, I’m sorry but I have no idea. Not even remotely familiar with Linux.

1 Like