[CrackMonkey] [email@example.com: [Pigdog] SVFIG presentations by Dr. Ting and Wil Baden]
nick at zork.net
Mon Oct 30 09:13:48 PST 2000
The nefarious Dr. Ting and his henchman, Mr. Baden are up to
their no-good FORTH plots again! This time, they are performing
unholy experiments to generate a FORTH engine in hardware with FOUR
STACKS! Will our hero be able to stop them in time?
----- Forwarded message from Jeff Fox <fox at ultratechnology.com> -----
Dear MISC readers:
Yesterday at the Silicon Valley Forth Interest Group meeting Dr. C.H.
Ting gave a presentation about the development that he is doing in
Taiwan with a new startup. They have been looking at ways to extend
his P16 and P24 designs. With five bit opcodes they have 32
instructions and Chuck had been using 27 in his designs so they had
five instructions to play with. And the 24 bit design can support 4
slots of 6 bit instructions or 64 instructions to play with.
They have looked at variations in the architecture and considered
several architectural extensions. The P16 offered the A register as
an A register stack with no change to the instruction set. A! pushes
the previous value in the A register into the Astack and the A (A@)
instruction pops an old value when it transfers A to T.
This means P16 is really a 3 stack machine and they have looked at 4
stack variations as well. This involves looking at the way it would
be represented in schematic or VHDL format as well as considering what
changes it implies in the instruction set.
With many new instructions to choose from they had considered a
variety and Dr. Ting explained the reasoning. He covered things like
short literals, branching, and math.
He also described a machine with 3 stacks in which the third stack was
3 registers wide X Y Z and which was synchonized to the movement of
the Return stack when CALL and RET instructions are executed. This
local stack would not be moved by >R and R> instructions. Dr. Ting
described variations in the way the local registers could be
initialized and used. He had coded an eForth for each of the designs
and talked about some of the more complex words in eForth and how the
Local registers were used and how one of these could contain the count
for the TIMES opcode to count down. He showed how this was used in
the multiply step instruction which combined the +* and 2/ operations
in the MISC instruction set.
In the afternoon session Wil Baden gave a presentation where he
reviewed the contents of Neil Bawd's "ugly web page" and explained
some of the details of his development style, how some of the tools
work, and how these are related. We will be treated to another
presentation at the Forth Day meeting in November in two weeks.
I was able to get videos of Dr. Ting's and Wil Baden's presentations
and will make them available. I have also converted some videos to
mpg format and will distribute them on CD-ROM. I placed a greatly
compressed streaming video version of Chuck's interview from tape 1 at
my web site. It is in a format for dsl speed access and 28MB so I
don't think you want to try to play it over a slow dial-up modem. I
can place some small videos at the web site.
----- End forwarded message -----
CrackMonkey.Org - Non-sequitur arguments and ad-hominem personal attacks
LinuxCabal.Org - Co-location facilities and meeting space
Pigdog.Org - The Online Handbook for Bad People of the Future
You are not entitled to your opinions.
More information about the Crackmonkey