20 Jun
2005
20 Jun
'05
1:38 p.m.
This raises an interesting question: if you replaced a real PDP-11 that
people connected to online with an emulated one, would they be able to
know the difference?
--
If you were using SIMH you would, since it doesn't perform cycle
accurate simulation. Unless you ran it on fairly slow modern hardware
it would be faster than the original.
This wouldn't be the case for MESS, though, since it is based on
the cycle-accurate MAME processor cores.
20 Jun
20 Jun
2:26 p.m.
cycle-accurate MAME processor cores.
For a Unibus PDP-11, what is "cycle-accurate"? Memory access
waits until the memory says it's done. I suppose you could come
up with some numbers not too far off for some memory implementation
and build around that. (don't forget modify as well as read and
write timings!)
Tim.
23 Jun
23 Jun
2:44 a.m.
Tim Shoppa wrote:
Wouldn't it be hard to tell, if you were unsure how accurate the timing was?
How do you know your timing routine is actually responding within a
certain time?
I'm reminded of my late father testing a homebrewed frequency counter,
finding it to be inaccurate, using it to measure its own 1MHz source
(which of course read bang on 1MHz), then realising that this was silly
and he'd need to borrow another frequency counter from work to discover
that his 1MHz crystal seemed closer to 900kHz...
Gordon.
cycle-accurate
MAME processor cores.
For a Unibus PDP-11, what is "cycle-accurate"? Memory access
waits until the memory says it's done. I suppose you could come
up with some numbers not too far off for some memory implementation
and build around that. (don't forget modify as well as read and
write timings!)
Tim.
7289
days inactive
7292
days old
2 comments
3 participants
participants (3)
-
aek＠bitsavers.org -
gordon＠gjcp.net -
shoppa_classiccmp＠trailing-edge.com