The parallelism can happen on various layers of the statement execution and so it is often hard to predict how large it will be. SAP Note 2100040 provides some background information about CPU utilization in SAP HANA.
My favorite approach to understand the root cause of CPU spikes is SQL: "HANA_Threads_ThreadSamples_AggregationPerTimeSlice" (SAP Note 1969700). With a properly configured "Modification section" (e.g. TIME_SLICE_S = 60, a useful AGGREGATE_BY condition for your purpose and no filter restrictions) you can check for top users, SQL statements or connections being executed during the critical time frame. Often it becomes quickly clear which SQL statement (i.e. STATEMENT_HASH) or user is responsible.
