When the system can group your I/O requests when initiating device commands, you can get much better performance. In BSAM and BPAM, the first READ or WRITE instruction initiates a short device command unless the system is honoring your MULTACC specification in the DCBE macro. The system puts subsequent I/O requests (without an associated CHECK or WAIT instruction) in a queue. When the first I/O request completes normally, the system checks the queue for pending I/O requests and builds a channel program for as many of these requests as possible. The number of I/O requests that the system can group together is the maximum number of requests that the system can process in one I/O event. This limit is less than or equal to the NCP value. If you code a non-zero value for the MULTACC parameter, you give permission for the system to group I/O requests more efficiently.

Recommendation: Use the MULTACC and MULTSDN parameters in the DCBE macro for the maximum performance with BSAM, BPAM and QSAM.