I’m curious to learn how one would go about to compare and benchmark the performance of different control electronics via the Juice OS.
Is it recommended to use Quantify as agnostic HW abstraction layer, or is it also possible to use the native language, say QUA, qblox-scheduler or Lab1Q?
I also wonder how easy it is going to be to evaluate the data, e.g. execution time and achieved performance, say coherence times or single qubit RB, and whether there are tools to accurately capture and reset the system state to ensure the same conditions for different test runs.
While Quantify is the recommended HW layer, it is possible to use any other package to control your hardware. This is as simple as installing the package in your container which is used to control the hardware and adding hooks to the database. For viewing respective GUI from these packages, the user will also have to add entrypoints to the Juice dashboard. However, functionalities such as viewing runtime data (plotting monitor), monitoring of active Qcodes instruments and other features from Quantify integrated into Juice will not be available. To summarize - this is possible via Juice, you can also take a look at the documentation here.
Juice provides a way for users to log their data to a time-series database and provides functionality to access this data. Evaluation of data and writing analyses depends on how a user approaches the problem. However, we do provide Juice extensions that allow the users to use our proprietary tools, our quantum diagnostic libraries that evaluate experiment data and provide performance metrics like you had asked for
