26 May
2016
26 May
'16
6:51 a.m.
On May 25, 2016, at 10:58 PM, Chuck Guzis <cclist
at sydex.com> wrote:
On 05/25/2016 07:22 PM, Paul Berger wrote:
Apart from that, it's not credible for another reason. CDC Cyber operating systems
always spooled printer output to disk (unlike OS/360 which did it in some variants but not
others -- notably not OS/360 PCP which I used since our 360/44 wasn't big enough to do
better). So a call to DMP would run only long enough do perform the formatting of
whatever memory was being dumped, writing the resulting text to the disk file named
"OUTPUT" for the invoking process ("control point").
paul
Speaking of dumps I remember an engineer friend
telling me that at
the university that he went to they had a CDC Cyber system and they
discovered that you could initiate a dump from any workstation, and
the system would dump out to the printer and while it was dumping the
whole system came to a halt.... guess what the students where fond
of doing...... It would seem to me that something like that should
have been more restricted.
At least in SCOPE and KRONOS, DMP was the command to dump memory--but if
initiated from a user's control point, it would dump only the user's FL,
not the whole system. So the story seems to be a bit apocryphal to me.
Most university systems charged not only by the CPU second, but also
by the number of lines printed and the number of cards punched.