Mike wrote...
Here's a quick chain test for doing 255 (+-)
chains with timing for my
SIMH configuration as described (about 23 sec.) I would hope that real
hardware
would out perform it. If not I'll have to think
about what could be wrong
with
the simulator and come up with another test. Let me
know what you get,
Jay, next
time you have her up.
I should have some time to run the chain test sometime this weekend, doesn't
appear to be much planned on the honey-do list. I'll let you know how that
test runs.
In the future if you want me to run anything to compare against the real
hardware, just let me know. I would think an emulator would run circles
around the real hardware though, at least hardware of the 2100 vintage. One
suggestion. I think you mentioned some sluggishness in the login procedure
under SIMH, and you also mentioned the SIMH approach was tieing the two cpus
together with sockets. I was using sockets for part of HPEMU, and found them
to be dismally slow, at least comparatively for the transfer speeds I
needed. I reworked that part with shared memory and the speed problem in
that section is now thoroughly gone :) Just a thought anyways.
Regards,
Jay West