Patrick O'Keefe wrote:
> As I recall, some of the "registers" actually had hard-coded values;
> they could be used as base registers but little else. I have no idea
> what happened if you tried storing into them. That alone supports your
> "grossly incompatable" assertion.
> It didn't even bother to pretend its linkage conventions were the same as
> s/360. It used something other than the s/360 linkage instructions BAL
> and BALR (BAS and BASR, as I recall).
>
>
that has nothing to do with the s/360 machine architecture. There are
differences between DOS/VSE and MVS as far as the linkage conventions
are concerned.
Looking at typical commercial applications probably 80-90% of the
instructions behaved identically.
In converting programs among the items that caused concern:
1. The pseudo registers. Many of our programs automatically used 0-7.
Obviously in a DOS/VS or MVS environment conversion it wasn't possible
to programatically (sic) use these registers. In addition instructions
that became available, eg EDMK and TRT added further restrictions.
2. It was a 16 bit based machine, no fullwords and indeed no Load
Address instruction, however in a conversion exercise you just changed
LH reg,=Y(address)
to
L reg,=A(address)
Didn't substitute the Load Address in case (address) was an external
reference.
These were probably the major conversion considerations in regard to
differences between S260/20 and S370/125.
Most of our conversion problems arose from the fact that we didn't use
the provided macros eg. PUT GET etc because they caused the assemblies
to take 10->100 times as long.
And I reiterate it was S/360 because IBM said it was.
Ken
----------------------------------------------------------------------
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
