If you really want to try using real addresses, I STRONGLY urge you to use the DATOFF system service rather than using your own code. If all you are moving is one byte, then you are guaranteed that the source byte is in the same real 4K frame as the target byte, but if the length is >1 then you may overlay some part of the wrong target page (worst case) or move some bytes from the wrong source page (best case). You will need considerable code to ensure that such moves can never happen with your one and only one MVCL. Unless you have allocated your virtual storage in such a way that the real storage backing the virtual storage is also contiguous, then you will have to loop through the data and never do the MVCL with parameters that would cause such an error. I assume that's how DATOFF does it. If you can do the following, I would also suggest you run a test case of moving 8K bytes using DATOFF and while tracing each instruction by running your MVS under VM (or on your ow! n P390 or similar machine that you can run in standalone mode with instruction stepping enabled). Then you will understand the necessary overhead if you want to move data using real addresses.
Bill Fairchild Software Developer Rocket Software 275 Grove Street * Newton, MA 02466-2272 * USA Tel: +1.617.614.4503 * Mobile: +1.508.341.1715 Email: [email protected] Web: www.rocketsoftware.com -----Original Message----- From: IBM Mainframe Discussion List [mailto:[email protected]] On Behalf Of Dave Day Sent: Wednesday, March 31, 2010 4:50 PM To: [email protected] Subject: DATOFF Nucleus routines Have code that moves data from a known area in one address space to a corresponding area in another address space. This code executes multiple times per second. Code uses AR's and a MovePage instruction to get the data from one to the other, one page at a time. Since I know the real address location of the source and target of the move, I'm wondering how much overhead would be saved by doing this using real addresses as opposed to virtual, and using one MVCL instruction. Anyone have any history with this? --Dave ---------------------------------------------------------------------- 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 ---------------------------------------------------------------------- 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
