Recursion error on selective machines

[Henry Wang]

Hi!

I've got a DGP problem which runs well on my local machine. However, when I transpose it to a computing node, I run into "RecursionError: maximum recursion depth exceeded while calling a Python object".

The issue here is that I need to embed the this optimization solver within an ML framework on GPU, and my local machine does not have a GPU (or any fast way to perform gradient descent).

I saw the response on "https://groups.google.com/g/cvxpy/c/btQuh4FsQ-I", however this solution didn't work for me. Specifically, *variable was returning a TypeError.

For more info: 

local machine - core i5-6200U; 8 GB RAM

node - intel xeon gold 6142; 16 GB RAM

Thank you!

Henry

[Steven Diamond]

I don't have a good answer, sorry. Are you using cvxpylayers? You could post an issue on their github: https://github.com/cvxgrp/cvxpylayers

[Henry Wang]

Hi Steven,

Thanks for the tip. This looks pretty cool - I actually wasn't aware of cvxpylayers before, and it looks useful.

I ended up implementing a fix by Harris Teague who suggested to just increase the max recursive iterations via 

"import sys

sys.setrecursionlimit(10000)".

Best,

Henry