• @mmcky to setup GitHub custom runner on QuantEcon Server

@HumphreyYang , if you adjust as per #238, the merge conflict should be fixed.

@HumphreyYang once main is merged into this PR -- the preview build should now get access to a GPU running on an EC2 instance.

@jstac and @HumphreyYang I have disabled the build cache and enabled upload of execution reports upon failure to get a clean run and see what the build issues might be.

I am getting a type error when running this local (but that is on an M1 OS X machine)

TypeErrorESC[0m: '<class 'jaxlib.xla_extension.DeviceArray'>' object does not support item assignment. JAX arrays are immutable. Instead of ``x[idx] = y``, use ``x = x.at[idx].set(y)`` or another .at[] method: https://jax.readthedocs.io/en/latest/_autosummary/jax.numpy.ndarray.at.html
TypeError: '<class 'jaxlib.xla_extension.DeviceArray'>' object does not support item assignment. JAX arrays are immutable. Instead of ``x[idx] = y``, use ``x = x.at[idx].set(y)`` or another .at[] method: https://jax.readthedocs.io/en/latest/_autosummary/jax.numpy.ndarray.at.html

I am getting a type error when running this local (but that is on an M1 OS X machine)

I think that is intentionally kept to show the difference between numpy and JAX. I guess the error comes from line 190 which is intentional.

thanks @Smit-create added a raises-exception tag.

@HumphreyYang @Smit-create I am getting a missing variable error

NameErrorESC[0m: name 'β' is not defined
NameError: name 'β' is not defined

on compute_call_price_jax.

I've timed out just now -- but will try get back to this tonight.

@mmcky, I was fixing that at the same time. It should be working after adding the condition.

@jstac the latest error I am getting seems to be a missing component from jaxlib

cannot import name 'pocketfft' from 'jaxlib' (/__w/lecture-python-programming.myst/lecture-python-programming.myst/3/envs/quantecon/lib/python3.9/site-packages/jaxlib/__init__.py

How do you install jax and jaxlib? I have tried pip but perhaps conda is a better option?

• merge ENH: GPU Support for Preview, Cache and Publish #243 and then update this PR from the main branch

🚀 Deployed on https://636193628e0a605e786ce814--epic-agnesi-957267.netlify.app

🚀 Deployed on https://636194245bb9625b42171f52--epic-agnesi-957267.netlify.app

🚀 Deployed on https://6361b881d516481100ecdbf6--epic-agnesi-957267.netlify.app

@jstac this PR now only contains the JAX intro lecture. Thanks for letting me use this as a test case -- greatly appreciated. All infrastructure changes are now on main

🚀 Deployed on https://6362040a60bdc93cc2c0d557--epic-agnesi-957267.netlify.app

@mmcky Please update us on the status of this PR when you have time.

@jstac this PR only contains content now. So if you're happy with the lecture we can publish it when you're ready.

I merged in #243 with all the testing we did on this PR.

Thanks @mmcky .

I wonder if we should be indicating to readers that what they see is output of a machine connected to a GPU, and hence what they produce at home might differ...

Anyway, please merge when ready. Let's get this live and then iterate.

nice thinking @jstac just building in a note on the jax lecture and updated the hardware information that is on the status.md page.

🚀 Deployed on https://637351564587ec23e2992248--epic-agnesi-957267.netlify.app

🚀 Deployed on https://6373518b09b24c1f85bc43be--epic-agnesi-957267.netlify.app

🚀 Deployed on https://6373569eef6bb5218f6ddb71--epic-agnesi-957267.netlify.app