David Ittah

> I have just done this test in Linux Ubuntu and it went fine: You're baseline already succeeded though, right? The issue here might not be related to dependency management,...

> The reason I opted for this design was to easily replace Pauli gates with calls to the Pauli frame tracker. Replacing (or erasing) a quantum gate is actually quite...

> Interestingly, when I run this locally I no longer see a difference between the first-call runtime and subsequent-call runtimes. It's possible that when `timeit` was executing these commands in...

Thanks @joeycarter, nice digging! So the overhead is not related to catalyst, happy to leave it at that and not show 1st vs 2nd call 👍

> > @joeycarter oh, actually this plot is quite nice! I would be happy to keep it in informationally, but we would need to change the text in the demo....

> > I agree with the framing of the compilation / runtime tradeoff that is typical for JIT, but note that the program in the demo is not parametrized at...

Nice PR! 🤩 > Pipeline stage renaming: Renames the first compilation stage from enforce-runtime-invariants-pipeline to user-transform-pipeline for better clarity. This last point is a bit questionable, because the pipeline's purpose...

> Is there a concrete reason for preferring the user passes to be applied in the QuantumCompilationPass as mentioned in the Slack thread? As far as modifying quantum ops goes,...

> I completely agree that unless there is a specific reason, all of the quantum passes should be run at the same stage. whether we want to apply them before...

> I think it is ok. > > I wouldn't do it at this level of abstraction though. When tracing gphase and extract you can use `jax.numpy.promote_types` to find the...