Synchronized MRP/DRP Calculations
On SMP computers, you can use the Synchronized Calculation and Synchronization Code fields to run multiple MRP/DRP sessions at the same time. On these machines, you can run one session on each available machine for a given synchronization code against a single database. Without SMP computers, these fields have no impact on DRP or MRP processing.
To run synchronized MRP/DRP, use any of the materials or distribution requirements planning programs. Open multiple views of the program. In each view, specify selection criteria for site, item, and so on. Set Synchronized Calculation to Yes and enter the same Synchronization Code for all views. Then start the program in each view.
During synchronized runs, the system creates a master work table listing all records that require planning and then updates the low-level codes. Then each session selects the first available record from the work table, planning the lowest-level items first. When an item is planned, its MRP Required field is set to No, so it is not replanned by another session. When a session completes planning at the lowest level, it waits until the other sessions are complete at that level. Then all sessions plan the next level up. This continues until all items are planned.
For more efficient processing, start the sessions with identical selection criteria—for example, the same site range—then all sessions select the next available record for planning. This results in the most efficient processing. When any planning level is finished, a session only waits for the other sessions to complete that level before all sessions can start on the next level.
If you start one session for each site, or for distinct site ranges, then when a session completes a level, it must wait for all other sessions to complete planning at that level. For sites with many items at a given level, that creates many more orders, requiring more time to process. Even if one SMP machine is being used by another unrelated process, you could end up with temporarily idle sessions. The advantage of running sessions by distinct ranges is that the audit reports print the planned items in the ranges you entered in each session.
Synchronization codes distinguish a synchronized calculation. All sessions started with the same synchronization code access the same work table of items for replanning. If a session is started with a different synchronization code, it will run without waiting for other sessions running at the same time. This can cause locking problems and double processing. A blank synchronization code is valid.
In some situations, running one session in non-synchronized mode may increase processing efficiency.
Example: You have a DRP receiving site that generates demand to two other sites but receives no demand itself. You run the session for the DRP receiving site in non-synchronized mode. That site is planned without waiting for the supply sites, and the two DRP supply sites wait only for the first site to generate the appropriate DRP demand. Once this is created, the two supply site sessions start their runs without waiting for the entire first calculation to complete.
Once again, it would be more efficient to run as three synchronized calculations for the same site range. However, using separate site ranges would make the audit reports simpler to read.