>how much overhead would be saved by doing this using real addresses as opposed to virtual, none, even if you didn't have the extra overhead of setting up to use, and branching to, the DATOFF MVCL function in order to do a move using real addresses (and even if your virtual storage happened to be backed by contiguous real).
>and using one MVCL instruction You could have done your AR-mode stuff with one MVCL instruction too if you wanted. If you're comparing MVCL to MVPG, the odds are good that MVPG wins. If you're comparing a single long MVCL to multiple MVPG's, I'd still guess that MVPG wins, but I defer to the expertise of others on the matter of instruction performance. Peter Relson z/OS Core Technology Design ---------------------------------------------------------------------- For IBM-MAIN subscribe / signoff / archive access instructions, send email to [email protected] with the message: GET IBM-MAIN INFO Search the archives at http://bama.ua.edu/archives/ibm-main.html
