On Thu, Apr 7, 2016 at 7:28 AM, Elardus Engelbrecht <[email protected]> wrote: >>Of course the computation(*) power per CPU is also growing significantly > => total power of SAP is growing even more. >>What is the rationale behind ? > > Good question. After some reading, I think it is to spread out I/O workload > from the CPU, otherwise all of them havee to wait for SAP? Oh, I could be > wrong of course. > > Groete / Greetings > Elardus Engelbrecht > > [1] - I see this on big blue's knowledge centre this interesting snippet: > > "Every modern mainframe has at least one SAP; larger systems may have > several. The SAPs execute internal code to provide the I/O subsystem. An SAP, > for example, translates device numbers and real addresses of channel path > identifiers (CHPIDs), control unit addresses, and device numbers. It manages > multiple paths to control units and performs error recovery for temporary > errors. Operating systems and applications cannot detect SAPs, and SAPs do > not use any "normal" memory." > > But something bothers me now - can anyone read above snippet and define > 'larger systems'? > z900/z800 had a max of 16 CPUs. z13(s) is over 100. Large would definitely be multi book mainframes. Of course even single book mainframes have been getting quite fast and large.
-- Mike A Schwab, Springfield IL USA Where do Forest Rangers go to get away from it all? ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: INFO IBM-MAIN
