Does MVS (DFSMS, whatever) still do that? What we used to call "double buffering"? I thought they had gone to a scheme in which they used one SIO to fill all of your buffers, while you waited. The theory being that CPU/IO overlap would be between your task and other tasks/jobs?
Am I wrong on that? Does MVS divide your BUFNO in two and fill one set while you process the other? Charles -----Original Message----- From: IBM Mainframe Discussion List [mailto:[EMAIL PROTECTED] On Behalf Of Ron and Jenny Hawkins Sent: Friday, August 26, 2005 7:18 PM To: IBM-MAIN@BAMA.UA.EDU Subject: Re: Most effective BUFNO? Charles, This is why in my earlier post I said that 16 works for half track blocking. 8x27998=223984 which means the chain length will be limited to 8 half track blocks. However MVS loves CPU/IO overlap, so we have another eight buffers that can be processed while the other eight are used for IO. I like Barry's formula, because it is not arbitrary - it is based on the limits of SAM-E. However I would recommend doubling his result. ---------------------------------------------------------------------- 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
