22 Oct
2006
22 Oct
'06
1:14 a.m.
On Saturday 21 October 2006 20:39, Jay West wrote:
That works well for many things, but not all. Bob, and the rest of the
contributors including myself, don't simulate the entirety of the
machine or peripherals such that diagnostics will necessarily
successfully run. As a concrete example, memory CSRs are not simulated.
Some diagnostics need this to work correctly so that they can set the
wrong parity, etc. Cache is another feature of some of the models that
is not simulated at all. I actually have SIMH modules for memory
controllers (including CSRs) and the J-11 cache (but not other models)
that I will submit for inclusion in SIMH someday.
Nonetheless, Pat is right. SIMH (and E11, I imagine) can be an
excellent reference platform for much machine behavior.
John
Tim wrote...
I'd feel better if someone having a similarly configured 34 could run
the same diagnostic and see if they get the same odd result :) If so,
then I wouldn't mistrust my machine quite so much.
You could always try using simh as your "reference" machine... :)