27 Nov
2006
27 Nov
'06
2:40 p.m.
Gordon JC Pearce wrote:
The thought occurred to me that it might be nice to
have the
floating-point option in my PDP-11/73. Since I haven't got one, and
don't fancy paying the $500 or so that some online sellers seem to be
asking (do they really expect to sell *any*, never mind at that
price?), I was wondering how easy it would be to implement in an FPGA
as was discussed at length earlier in the week.
Feasible, or a complete waste of time? Presumably I'd need software
written to make use of the floating-point hardware anyway...
Jerome Fine replies:
You have not stated your criteria as to why you want the faster speed.
(a) If your ONLY reason is faster execution, then use Eratz-11 and get
up to 100 times the speed of a real DEC PDP-11/73 on a Pentium 4 CPU.
Since I am able to achieve about 20 times the speed of a real DEC PDP-11/73
with a Pentium III, I suggest that 100 times the speed with a Pentium 4
is a reasonable estimate.
(b) If you want faster execution, but it must be on a real DEC PDP-11/73,
then you have 2 conflicting reasons. And since you can already execute
the FPU instructions, but slower without the extra hardware option, it
seems reasonable to assume that if you want the real DEC PDP-11/73 CPU,
you do not really care that much about the overall speed of the CPU.
Can you please clarify?
I also have a PDP-11/83 (which always included the floating-point
option). In addition, PMI memory improves the overall speed by about
another 20%. I am much more focused on the software, so I really
don't prefer any particular hardware and I now rarely use the DEC
hardware except to test for any actual DEC hardware conflicts.
Sincerely yours,
Jerome Fine
--
If you attempted to send a reply and the original e-mail
address has been discontinued due a high volume of junk
e-mail, then the semi-permanent e-mail address can be
obtained by replacing the four characters preceding the
'at' with the four digits of the current year.