ken@rochester.UUCP (Ipse dixit) (12/15/85)
is in
bankruptcy, thus there is not much support for the product. I don't
know who is buying up all the assets, but there were several bidders.
They might still be listed in San Diego. Their system runs on PCs
and the rest of the P-system machines. There is a special hack to
run P-system programs like ordinary MS-DOS programs. Hmmm. You might
check with Stride Micro in Reno, NV (formerly, Sage Computers). They
may be selling a native code P-system compiler as well as the standard
Volition release on their 68K based box. If you want more details,
I can get them.
Rob
------------------------------
Date: Tue, 10 Dec 85 23:07:52 est
From: treid@mitre.arpa (Thomas Reid)
Subject: Re: p-code compiler
To: broman%bugs@nosc.ARPA, info-modula-2
I hope someone can be more exact, but I remember the Volition Modula-2
product being hosted on top of the p-system.
------------------------------
Date: Tue 10 Dec 85 23:27:48-EST
From: Ralph W. Hyre Jr. <Ralph.Hyre@C.CS.CMU.EDU>
Subject: Re: p-code compiler
To: treid@MITRE.ARPA
I called Pecan Software Systems today. (Pecan is the company that acquired
Softech MicroSystems' assets, which is the company that acquired some of
Volition Systems' assets, which is the company that developed the P-system
based Modula-2 compiler. Which means Pecan might have it.)
They told me to check back in about three weeks for more information on
products, but they didn't specifically say they were going to introduce a
Modula-2 compiler based on the P-system. I suspect they'll be busy getting
the vanilla P-system products re-introduced, if anything.
One thing they did say was that they weren't planning to sell a native code
generator which would compile the stack machine-like P-code instructions into
6502 assembly language instructions (for an Apple)) [Given the maturity of
Apple Pascal, it seems unusual that NOBODY has done this, does anyone know
if a 6502 native-code generator has been written for the P-machine?]
Needless to say, I'll report back in three weeks.
- Ralph Hyre
- -------
------------------------------
Date: Wed, 11 Dec 85 08:29:35 pst
From: Vincent Broman <broman%bugs@nosc.ARPA>
Subject: Re: p-code compiler
To: info-modula-2
i was able to phone volition systems yesterday. i never got any sales people,
but talked to a software auditor. he said that pecan was buying the modula2
compiler from volition systems, but the deal was contingent on his examining
the software and certifying it worked, since the authors have already headed
for the four winds. so, technically, volition still owns the stuff and they
might be doing business till everything is officially handed over.
this fellow (who seemed to be some kind of contract programmer) already had
the opinion the compiler worked - in fact he was quite impressed with its
performance.
modula2 runs under a ucsd p-system, but compiled modula2 code and (current)
p-system ver IV.x code represent slightly divergent extensions of ver II.x .
they are both probably backward compatible, so each could probably
run old ver II.x code, but not cross compatible. he opined that pecan would
probably rework the modula2 system to harmonize with the current p-system,
since pecan "is so big on ver IV.x". (pecan is already advertising it in
infoworld.) he said one incompatibility was the "dynamic linking" of modules
in their modula2 -- which point i didn't savvy.
i asked about native code generation. he said they had none, but that stride
micro had some p-code to native mc68000 code generation in their compiler,
which is derived from volition system's compiler.
when i inquired about the ms-dos implementation, he said that the distribution
included modules for dos calls, peek&poke, etc.
i'll continue to try to contact sales people at volition systems. i imagine
a p-code implementation will be fine for my little home system, since
the floppy disks will make most things io-bound, anyway. and i would
appreciate the nice integrated development environment. the other
compiler i'm considering is logitech v2.0: US$89 for compiler+editor.
i like its native code generation, interrupt raising and handling,
register refs and dos calls, etc. BUT, i understand that the 4pass compiler
is very slow. anyone else venture a judgement?
Vincent Broman MILNET: broman@nosc
Analysis Branch, code 632 UUCP: {ihnp4,decvax,akgua,dcdwest,
Naval Ocean Systems Center allegra,ucbvax}!sdcsvax!noscvax!broman
San Diego, CA 92152 Phone: (619) 225-2365
------------------------------
Date: Wed 11 Dec 85 11:57:06-EST
From: Ralph W. Hyre Jr. <Ralph.Hyre@C.CS.CMU.EDU>
Subject: Looking for m-code specs and/or m-code interpreter
To: info-modula-2
This is probably something I could get if I had a Lilith, but I'm hoping
it's available somewhere else.
I also need an m-code interpreter that is written in one of the following
languages (listed in order of preference):
Any high-level language (but preferably Pascal or C)
6502 assembly language
p-code
pseudocode or assembler for any general-register machine
Please reply to me directly and I will summarize to the list.
- -------
------------------------------
Date: Wed, 11 Dec 85 12:32:02 CST
From: frank@db.wisc.edu (Dan Frank)
Subject: Re: p-code compiler
To: broman%bugs@nosc.ARPA
I bought the Logitech compiler back when you had to pay real money for it
(v 1.0). Each version has become steadily faster, although it doesn't win any
awards in that department. I will say that back when I had an XT, I developed
some sizable multi-process simulations using v.1.1, and compile/link time was
not the problem; run-time process switch overhead was (to be fair, I ported
the simulation to an 11/750, and it was pretty slow there, too).
The version 2.0 compiler is said to be somewhat faster than 1.1, and if you
get the non-overlayed version (requires a 512K machine), it should be signifi-
cantly faster. You can also get a syntax-directed editor that ties the whole
system into a Turbo-like integrated environment.
I am waiting for my 2.0 upgrade. If I have any further insights after I
receive it, I'll post them here. As it is, though, I would strongly recom-
end it to anyone with a PC, especially at the new prices.
-- Dan
p.s. The run time debugger is worth its weight in gold.
------------------------------
Date: 11 Dec 1985 10:23-PST
From: Guy Tiphane <logitech!guy@decwrl.DEC.COM>
Subject: re: p-code compiler and logitech v2.0
To: info-modula-2
I am a user of Logitech V2.0 (I got the release before it was sold,
since the developers are in my building!), and found that it is
significantly faster than the previous version. There is no official
benchmark published, but here is how I see it:
Compiler: V2.0 looks like it's at least 33% faster
(i.e. it takes two thirds the time it used to)
Linker: V2.0 looks like it's around 50% faster
(i.e. it takes half the time it used to)
Exec: I can't really tell at this moment, due to the
nature of my applications
I use an AT&T 6300 with 640K and 10MB HD...
------------------------------
Date: 11 Dec 85 22:27:00 EST (Wed)
From: Bob Munck <munck@mitre-bedford.ARPA>
Subject: Logitech 1.1 Compiler Speed
To: Info-Modula-2
I'm running the old version on an 18 MHz PC-AT with all libraries and
code on a 3MB RAM disk. It's faster than all Heck!! My application uses
the MetaWindows package from MetaGraphics and the 80287 chip, and it's
also impressively fast. I've cobbled together a syntax-directed editor
using MINCE and PROKEY, an archive and configuration management system
using ARC and command procedures, and a system to trace inter-module
procedure and data references using utilities like grep and sed from
local bboards. It's the most productive development system I've ever
worked on, and I've been in the business of development systems since
1970.
So how do I get this new release?
-- Bob Munck, MITRE
------------------------------
Date: Wed 11 Dec 85 21:55:19-CST
From: Pete Galvin <CC.GALVIN@R20.UTEXAS.EDU>
Subject: Logitech 2.0 educational discount
To: info-modula-2
I recently called Logitech to find out whether there'd be a discount given
on the new (cheaper) version of Modula-2. The salesperson I spoke to said
that for $125, all of the modules listed in the Byte add can be had.
This isn't as big a discount as they previously gave, since the list price
of the set is only around $225. Still, for $125 you get the compiler, editor,
symbolic debugger, math support and utilities. My order goes in soon...
(since they weren't yet shipping when I called, I decided to wait a little.
I forgot to ask how I should show I was "educational". Any guesses?
The salesman also mentioned that I should spread the word about the discount...
so I am.
--Pete
- -------
------------------------------
Date: Wed, 11 Dec 85 15:29:32 PST
From: sun!orb!nagler@ucbvax.berkeley.edu (Rob Nagler)
Subject: Re: p-code compiler
To: broman%bugs@nosc.arpa
I have used both the Volition and Logitech systems (the latter quite
extensively). Here some of my observations:
Logitech:
1. Static linking causes problems with Program.Call (subprogram execution).
2. Code generator is not that brilliant, but it is fairly reliable.
3. Supports XXX87 math coprocessor as well as an emulator, so you don't
need one to do floating point.
4. There coroutines are slow, but you can write your own (IO)TRANSFER
very easily which is a lot faster (they may have improved the
performance in their latest release).
5. The Debugger(s) are great (especially the Postmortem one, cause you
can retrace after the fact instead of having to start up in it).
6. Each module has several files associated with it (DEF, MOD, REF, SYM,
and LNK), so there is a significant file overhead. If you only have
a floppy, this can be slow.
7. It has been run through the Modula-2 Validation Suite (MVS) and it
faired well (after a few bug reports were fixed). Given that it is
a derivative of the Lilith, it had better be close (although you would
have been very surprised about the bugs in the earlier releases).
Volition P-code:
1. Dynamic linking is wonderful! You don't have to have MAP files and
you don't need a linker (believe me it is much faster).
2. The module libraries are kind of awkward to use (you have to have a
unique number for each of your modules in a program).
3. It doesn't compile real M2, that is, you can't have forward references
(only one pass is faster compilation, but you have to use "forward").
4. I found the floating point support atrocious. It totally wedged my
Compaq when I did some fairly simple stuff.
5. The P-system is in general a win on tiny systems, because it generates
compact code and compiles real fast. Obviously, you aren't going to
get a lot of performance during the run time.
Volition 68K Native (only close second hand experience):
1. Very high quality code output (separate optional optimization pass).
2. Still no forward references.
3. Still dynamic linking (yeah! yippee! somebody did something right.)
4. In general, the people I have talked to who have used it, are quite
happy with it.
Personal opinions:
Logitech: good for people who want to develop real-time code or
actual commercial products on the PC. In other words,
you can build big systems on the PC very easily.
Volition P-code: good for people who need quick hacks, like to
play around, better for large scale programs than UCSD Pascal,
portability is obviously there, etc.
Volition 68K Native: unless you have a Scenic or Stride (Sage), I would
not suggest buying it (unless you can get source). The people
at Stride are very happyy with it and they are real performance
conscious. If you are not at a University, this probably
(is currently) your only alternative for a quality 68K
compiler.
Rob
[Everything I in this note is personal opinion and does not represent the
views of my employer or his friends. Please don't sue me. I don't have
a lot of money and I don't really like talking to lawyers.]
PS. Rumor of the month: The MVS has made it out of the lawyers hands at
Hayes and the final paper work was sent to the British Standards
Institute (BSI). This means that you compiler developers may actually
get to use it (wonders of wonders).
PPS. Sun Microsystems (my employer) is working on a M2 compiler which
is a derivative of the Lilith System. The rumor is that we are
going into a beta release sometime around February. If you want
to be on the beta list or just want some information, contact
your SMI sales rep.
------------------------------
End of Info-Modula-2 Digest
***************************
--
UUCP: ..!{allegra,decvax,seismo}!rochester!ken ARPA: ken@rochester.arpa
USnail: Dept. of Comp. Sci., U. of Rochester, NY 14627. Voice: Ken!