On Oct 13, 2005, at 4:00 AM, Randy Hudson wrote:
In article <[EMAIL PROTECTED]>,
Ed Gould <[EMAIL PROTECTED]> wrote:
-------------------
SNIP--------------------------------------------------
I wonder if the CCW chain could be processed fast enough that much of
the
reversing could be done with the Chain Data flag and appropriate
destination
addresses. In the days when we actually wrote channel programs, we
could do
a lot with them; for example, a PCI could allow the program to start
validating the early data while the check was only partially read,
substantially widening the stacker-select window.
Randy,
I still remember in those days the POPS for each model had the " micro
seconds" each instruction took. My memory is sketchy and after the
document was read you had X (I don't remember) many microseconds to
make a decision as to which bin you could send the document to.
The POPS indicated how many ms each instruction took. Any time a
decision was to be made we had to calculate a total for the path length
so we didn't miss the window. Its been ages so I don't remember a lot
but I don't remember when you decided to send it to bin 0 (example) how
long it took for the selection to happen. I am sure it was documented
somewhere, I just don't remember as it was 30+ years ago.
Maybe someone with a better memory can pipe up with more explicit
detail than I can provide.
I still cannot believe that IBM would come up with a machine like that.
That was the only machine that I ever worked on that the timimg was so
critical.
Ed
----------------------------------------------------------------------
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
