ARTIQ + Labscript

[saesun kim]

Is it possible to integrate ARTIQ into the Labscript suite, considering ARTIQ has its own language? If anyone has done it before, can you share your experience or examples? Thanks!

[dihm....@gmail.com]

I've thought about this off and on a few times over the years.  At the surface level, it should be easy. ARTIQ has a python interface, just write a labscript device that translates labscript commands to ARTIQ ones and call it a day.  Ultimately, there have always been a few issues that stood in the way of actually doing it (besides time).

1) I've found it surprisingly difficult to actually get the ARTIQ hardware to test with. A solvable problem to be sure, but still annoying. The ethos of labscript is built on using readily available hardware.

2) The main thing ARTIQ brings to the table is in-shot logic/feedback at a hardware level. This keeps things fast and super flexible. Labscript doesn't really have a native way of accounting for that. I can imagine supporting branching shot trees in a single shot as complicated. Certainly possible, but probably tricky. Especially when trying to work with non-ARTIQ hardware at the same time. After all, there is a reason the ARTIQ hardware is a monolithic solution--to support this very feature.

3) Labscript's pseudoclocking logic/code is somewhat bespoke and a pretty core element. Trying to force ARTIQ into the pseudoclock paradigm could be tricky.

Every time I think about doing this, I end up in the same place. Once I have an experiment that actually needs in-shot feedback, it's probably best to just swap out labscript for ARTIQ entirely. Of course, I haven't actually used the ARTIQ front-end software so maybe that tune will change, but it's where I'm at with it. If you or anyone else is motivated enough to actually try this seriously, I'd be happy to help where I can though. It could be a very powerful addition labscript afterall.

-David