This includes inquiring if we have existing benchmarks for the new optimizer, how often the/any benchmarks are run, if we have anything we can leverage longterm to prevent/identify regression.

Also, let's try to identify our "embarrassingly parallel" use case.