Since I helped Barry with the tests in 1984 and have repeated it with every new DASD device and OS release since then (it is a nice test to look for changes) I will chime in. 240K/BLKSIZE and 30 are both hard limits for non- extended format QSAM files. Why 240K? I dont know. Why 30? Might have to do with it being a full cylinder. While the scale has changed some over the years as hardware has changed, the shape of the curves is essentially identical. The knee of the curve is still at 5 buffers for non-extended format QSAM. Extended format changes everything. The number of effective buffers there depends on how many stripes you have but there is still a knee to that curve. I have used as many as 240 but did not see enough improvement of 64 to justify the cost of the extra memory. Compression changes it some more since there is then an additional CPU factor to consider. YMMV, but more than 30 on non-extended format and more than about 64 on extended format are - while you can specify them - a waste of time.
---------------------------------------------------------------------- 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
