In a message dated 5/1/2006 1:38:21 P.M. Central Daylight Time, [EMAIL PROTECTED] writes:
>PAV is designed for multiple >accesses on different extents. 2105 PAV will allow up to 8 simultaneous reads of the same single track if none of those read commands is in a channel program that has signalled write intent. >Since all the jobs will be acessing the same >extent range, only one UCB will probably be used. As many UCBs will be used as WLM has defined as alternates/secondaries. You could easily have 8 UCBs on one system, or one UCB on 8 different sharing systems. >I like the idea of using HiperBatch the best. Work up a job to load the >dataset into hiperbatch buffers, then let the 35 jobs go at it. The world's most highly tuned DASD I/O request - a cached read hit on a short block - is still untold gazillions of times slower than doing no I/O at all by finding the data somewhere in virtual storage so a move instruction can be used. Bill Fairchild Plainfield, IL Gens nefaria Bushorum delenda est. ---------------------------------------------------------------------- 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
