No Subject
- To: MISC
- From: jfox@xxxxxxxxxx (Jeff Fox)
- Date: Sun, 11 Jun 1995 10:23:16 -0700
Dear MISC Readers,
Robert Hoeller writes:
>is there a way for me to get the MuP21 timing specs?
>
>I tried to order "More on Forth Engines" Vol. 19 from
>Dr. Ting but I didn't get any response from him. (Hello Dr. Ting!!)
>
>I would appreciate some electronic way of distribution.
>I don't think "information hiding" would help spreading a new design.
>
>Robert
I am curious, how did you attempt to order Vol 19?
Mail? Phone? Fax? Email? Did you contact Dr. Ting, or Offete
Enterprises Inc.? If you used phone which number? Dr. Ting
lists the phone number for Offete, but that is his father
who does the mail order business. Dr. Ting also has a home
number and a number at Applied Biosystems where he works. He
also has email there, but keeps a pretty low profile. I can understand
his not wanting to run the Offete mail order business out of his
employer's business office, using their phone or email for Offete
related stuff. Also Dr. Ting's email address changed a while back.
I am sure Dr. Ting is not trying to "hide" information about MuP21.
Please remember that Dr. Ting has had MuP21 developed and
manufactured, had two PCB for MuP21 made, has written various
software, has published 19 volumes of More on Forth Engines,
has worked on the FIG board, has been program chair for
the SVFIG, has done presentations almost every month for FIG,
has made trips to Taiwan, and has coordinated a lot of work
with Chuck and other people on this project. This has all been
in his "spare" time as he has a regular job to support his family
and finance this entire effort out of pocket. Other people have
contributed documentation, programs, explantions etc, and that is
why we have the MISC mail list. I am pleased that there are as many
people as there are helping to provide information and answer questions.
And I would encourage people to do more if they can.
Dr. Ting is not trying to hide information, or keep it
under key where only he can release it. Although it might
seem that way. You just cannot expect one guy to provide the
outlet of information on other stuff. I went to the largest
bookstore in the Western US recently, they had 350 books on
C and C++. The only thing on Forth was "Starting Forth." There
was no "Peter Norton's MuP21 GUI Guide" or "Inside the Norton
Utilities for MuP21".
Forth is obscure enough, let alone P21.
I think there are a few sources of any information on P21. Offete
has lots of printed material, and we have MISC and a couple of
www sites with information. That's it.
I think haveing MISC and the archives is great. I hope more poeple
will get involved in dialog, and organizational activities. I
think a FAQ, board related stuff, software related stuff,
online copies of technical stuff etc.
I will ask Dr. Ting what technical information he can donate to
MISC so that we don't have to retype PD information. I know he
is selling documentation and software as well as hardware, but
he realizes that he needs good publicity about the operation too.
Anyway forgive me if I have gone on and on. We are all subject to
valid criticism. I guess I just a little sensitive on this subject.
I may be overreacting to what I perceive as "the problem is that
_______ is not doing enough" type criticism. (insert Chuck, Dr. Ting,
or my name) Perhaps I am still reacting to thread back in c.l.f
a while back that I would call "the problem with P21 is that Jeff
is not a good salesman" (Two Chucks thread) You will notice that
I do not list "selling P21" in the list of things I am working
on. Anyway things get spread pretty thin sometimes, so we can't
always do as good of a job as we would like.
We will be in the same situation with F21, it may even be worse in the
sense that there is not as much funding. At the moment I am responsible
for getting technical information on F21, financing development, getting
chips made, reviewing designs, helping Chuck with testing the design
with the simulator in OKAD, writing network software, writing GUI software,
writing compiler software, writing parallelizing extensions, writing
test routines for prototypes, getting boards designed, getting boards
made, finding memory sources, writing documentation, doing presentations,
negociating with a long list of companies, administating the mail list,
and answering lots of who-what-when questions daily on the net.
**
( only a small percentage of time is spent in hardware simulation
tests, so a faster PC would result in negligable change in
productivity for myself or Chuck. Most of the time a faster PC
would be of no use because OK does most things in a fraction of
a second. The things that take time will still take time. It won't
matter that much if the PC simulation is 4 or 12 billions times slower
than the actual chip. There is a long list of things with much higher
priority than going from a 486 to a Pentium. But I do have to admit
I would enjoy SU-27 even more with a Pentium. :-)
____________________________________________________________________
I am posting the timing specifications that Dr. Ting put in the
More on Forth Engines Vol 19. This is clearly technical information
that Chuck put in the public domain, so I have no problem posting it
here. A lot of the information that Dr. Ting publishes is after all
public domain stuff. For people who are not on the net or want their
own printed copy Offete is there. But much of the stuff he sells is
available somewhere online.
I enourage other people to post technical information. Chuck tends to
give it to Dr. Ting or myself to release it into the public domain.
Dr. Ting publishes it in his newsletters etc, and I tend to put it
at my www site etc.
Something like the timing diagrams is not to be considered proprietary.
I would not publish the schematic to Dr. Ting's kit board without
permission, but this is not to say people should not fell free to
experiment with variations and publish the details.
Since I don't have a kit here I will be asking some questions about
the circuits on the kit board. I did have one here for a while, but
never did have a schematic or diagram, and the notes I did have had
errors that I reported to Dr. Ting.
MuP21 Part Timing Specifications Chuck Moore 1994
TC514400APL-80
RAS _______________
_| 60 |_________________________________________
CAS _________________________________ 20 ___________ _ __
| 60 |___| 30 |_____|
ADR ___________________ 10 _______________________________________________
|_|_|_|_|_|_|_|_|_|_____|_|_|_|_|_|_|_|_| 10 |_|_|_|_|_|_|_|_| |_|_|
WE _______________________________ 40 ____ _
| 40 |________| 10 |___________________|
Refresh
CAS__________________ 5 __
|_|____________________|
MF3512
SRAM __ 250
|_______________________________________________________________|___
WE _________ __________
| 30 |_________________________________________________| 30 |
TC514416P-15
FRAM __ 15
|___|___
WE ___ 15 ____
|___|
MuP21 SIGNALS
RAS _____________
__| 60 |________________________________________________
CAS __________________________________________ 30 _______ __
| 50 |_____| 25 |______|
ADR ___ 15 _______________________________________
|_|_______________________|__| |_|_|_|_|_|_| 4 |_|
WE _____________________________________ 40 ____ __
| 40 |_________| 10 |__________|
Refresh
CAS _____________ 20 ___
|_____|___________________|
SRAM __ 260
|______________________________________________________________|___
WE _____________ __________
| 30 |_____________________________________________| 30 |
FRAM ___ 40
|_______|____
WE ___ 40 ____
|_______|
_______________________________________________________________________________
You will note that F21 will use similar memory parts, but will also support
256kx4 DRAMS, 150ns and 250ns ROM, 12 or 25ns SRAM, and will have three bits
to change the overall memory timings. Memory timing will be more programmable
on F21 and will be harder to document. The memory setup times will be much
less than MuP21, partly because of .8 and partly because of more than one
power and ground pin.
Memory setup time is a critical thing. Last word from Chuck was the first
F21 prototype may not have the ideal memory setup, and will need a longer
memory setup time. If it is 3ns (as it can be ideally) you can get 15ns
throughput out of 12ns memories. If it is 8ns then you get 20ns which
would limit performance to 200mips. On MuP21 is is 25ns. This means
you get 40ns throughput from 15ns SRAM on P21.
It will interesting to see what the numbers will be for working F21 chips.
Jeff Fox
