Use of block sizes with APO pgm /SAPAPO/RDMCPPROCESS

[Former Member]

I am using APO V5.1.

I am using standard pgm /SAPAPP/RDMCPPROCESS to send APO planning data back to the connected ECC system.

One of the parameters is block size.

I am using the default of 1000.

But what are the performance implications of using different block sizes? What is 'good practice' for using this block size parameter?

Thanks for any advice on this point...

[former_member209769 · Accepted solution]

Hi,

The F1 help itself gives good information on this topic.

The block size indicates how many change pointers are processed in one shot (one block).

So, if you increase the block size, it means that more change pointers are processed in one shot meaning that the memory requirement would be higher. On the other hand, since you have increased the change pointers to be processed in one shot, the number of blocks required would be less. So, if your system has sufficient capacity to process the number of change pointers that you specify in a block, increasing the block size would increase the load on the memory, but would make the process fast, meaning that your change pointer job would finish faster.

Vice versa is also true.

Your system performance (mostly linked to memory availability) could get negatively impacted if you increase the block size. If you see that your system performance goes down with this block size of 1000, this should mostly mean that your memory resources are already exhausted in other processes, and you could reduce the block size to 500. With this, you free up memory and system would perform faster. Though your change pointer job would take a little bit more time.

Try first with having the default block size. Unless you face some issue, there's no point in changing the default value.

Hope this explains.

PS: In our case, we mostly use 1000, but during some peak system load times, we use 500 as block size.

Thanks - Pawan