Any good practice to setup cache for continuous integration runner?

[vallsv]

Hi,

At my company we are about to check conversion of our CI from CONDA+PIP into PIXI.

We have the full control of the gitlab runners.

For example we was using PIP_CACHE_DIR in order to reuse the cache between multiple jobs.

I just found there is also a PIXI_CACHE_DIR.

Do you know if there is good practice around PIXI in order to optimize continuous integration systems?

Thanks a lot.