coreyostrove
commented
6 days ago Commenting here to try to put fresh eyes on this proposal again. There have been a few instances recently over the past couple of months where I've wanted to build experiments for running DRB against native non-clifford gate sets and was unable to with out current implementation. This is mostly just a data point in favor of this being something we could and would make use of in practice if available.
The existing implementation of direct randomized benchmarking in pyGSTi has limitations regarding the need for the gate set to be able to be compiled into clifford operations. This feature is aimed at increasing the flexibility of pyGSTi's DRB implementation to more easily handle the construction of experiment designs for arbitrary universal gate sets. Development of this capability is currently underway on the branch drb-of-universal-gatesets.