In <[EMAIL PROTECTED]>, on 06/17/2005
at 11:20 PM, "Joel C. Ewing" <[EMAIL PROTECTED]> said:
>In the old days, when physical block sizes were small, and number of
>buffers used were typically small, and QSAM and DASD hardware wasn't
>nearly as smart, there were cases where special purpose I/O routines
>which either generated a channel program to read all blocks in a
>track or used hardware Read Track instructions could save both real
>time and CPU time.
Typically? It was always easier to increase the number of buffers and
to specify chained scheduling than it was to rewrite the program to
use EXCP. Yes, there were cases where you got a performance boost
using your own channel program, but it wasn't nearly as cut and dried
as what you suggest. As for Read Multiple Count Key Data, that wasn't
available in the early days; as I recall it came in with the 3330-11
and 3350.
--
Shmuel (Seymour J.) Metz, SysProg and JOAT
ISO position; see <http://patriot.net/~shmuel/resume/brief.html>
We don't care. We don't have to care, we're Congress.
(S877: The Shut up and Eat Your spam act of 2003)
----------------------------------------------------------------------
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
