I've been doing mainframe operations for about 4 years now. I'm continually amazed at the degree of control that a company has over their mainframes. I suspect this is why the suits like 'em.
Anyway, the reason some things run much faster than others on mainframes isn't usually owing to the power of "the CPU". It is usually because of dispatch priority and service class, etc. So, when you search your code, you are using the TSO address space, and you have a pretty high service class. When you submit your JCL to compile your program, it probably gets a much lower service class, and so finishes later. There are hundreds of programs, jobs, users on the system at any one time and they are all being controlled by several pieces of system software. Your job may even have to wait for an initiator to open up (another performance control)... Check around in SDSF to see what is going on.
Slashdot Mirror
I've been doing mainframe operations for about 4 years now. I'm continually amazed at the degree of control that a company has over their mainframes. I suspect this is why the suits like 'em.
Anyway, the reason some things run much faster than others on mainframes isn't usually owing to the power of "the CPU". It is usually because of dispatch priority and service class, etc. So, when you search your code, you are using the TSO address space, and you have a pretty high service class. When you submit your JCL to compile your program, it probably gets a much lower service class, and so finishes later. There are hundreds of programs, jobs, users on the system at any one time and they are all being controlled by several pieces of system software. Your job may even have to wait for an initiator to open up (another performance control)... Check around in SDSF to see what is going on.