4) In-order multi-threading What is the key difference between FGMT and coarse-grained multi-threading? What two performance metrics does FGMT trade off, in practice? What are the program behaviors that will make a collection of threads or a set of programs well- suited to an FGMT platform vs. an in-order superscalar processor. From an issue-bandwidth perspective, what limits the utilization of issue slots on an FGMT machine? a) b) c) d)