Eric, Thanks for the explanation, I get it now.
I was hoping to avoid the design decision that I have to make now--deciding beforehand what size of shared memory block to obtain. I had hoped that I could simply allow the available resources be available to me, not wasting any space with an overestimate, not getting thwarted by an underestimate. I am aware of RMM with SHM and will now implement my scheme using that setup. Your explanation of the spawning of child processes is what made all the difference. I see now how the shared memory works, I was misunderstanding it before. Thanks for your help and time. Rich Yonts sola fide, sola gratia, solus Christus, sola Scriptura, soli Deo gloria