Should "count" be included in distribution JSON output? It may be derived from data already present.

https://github.com/gem5/stats-schema/blob/main/simstats.schema.json#L83

There is no really good reason, in my mind, to include count when it's easily derivable from data already present. A function inside the Python stats library should definitely be included but it shouldn't be a separate piece of data we store.