Updates to job executions - optimizing classical computation

We have released a new feature that will parallelize some of the classical computation necessary to prepare a submitted job for its quantum computation. Till today all aspects of job processing were executed serially, meaning the target backend would be held from processing another job until its current job completes. This would be visible in your dashboard as having at most 1 job in the "Running" state at any one time. With parallel compilation, a user can see multiple jobs in the Running state and stay in the Running state longer than before. We have plans to provide more detailed job statuses in the near future, so users can see exactly where their jobs are. With this change, we also expect to see faster completion times for Qiskit Runtime® jobs. Currently, this optimization has been made available on ibmq_manila. We plan to deploy it on to the other systems in the coming weeks. It is important to note that, as a result of this change, a single Qiskit Runtime job will no longer have exclusive access to a backend like it's implied today.