I have a query which is joining 10+ tables to create another table. I will need to create multiple tables simultaneously by running the same queries, so each connection would read from these tables. And I notice that my performance is not gained linearly hence I do a profiling on this and notice that the CPU_USER value in the information_schema.profiling table is unreasonable large. Anyone could help to explain what is this CPU_USER and why it is large when I run multiple threads at the same time?
Thanks