February 2005 - Gambit-list - mailman.iro.umontreal.ca

[gambit-list] Suggestion for configuring gambit on high-end x86 processors.
by Bradley Lucier 16 Feb '05

16 Feb '05

Marc: High-end P4, Athlon, etc., as well as Athlon 64, have 64-bit floating-point instructions pretty much the same as any other RISC (I don't mean the fused multiply-add of IBM's Power). This is accessed through the SSE2 instruction set on Intel, and similar things on AMD's Athlon. Some algorithms in _num.scm (for converting strings to and from floating-point numbers, for example) rely on proper 64-bit floating-point arithmetic. Gambit simulates this by setting the rounding precision to 64-bits on x86's 80-bit floating-point stack, but there is still the possibility of double rounding, especially for denormalized numbers, because of the extended exponent range of 80-bit numbers on the floating-point stack. I'm thinking it might be a good thing for Gambit's configure script to test for the availability of SSE2 or the similar thing for AMD and to set those options when building Gambit with gcc. Brad

1 0

RE: [gambit-list] New types
by Logan, Patrick D 14 Feb '05

14 Feb '05

A functor is I think just a first-class function, either named or anonymous, e.g. a (lambda (...) ...) I think the usage here is more like "combinator", i.e. a functor with no free variables that can be combined with other combinators to create new kinds of compositions the way objects in an object-oriented program might be composed. (aka the "part-of" relationship rather than the "is-a" relationship.) I may be putting words in David's mouth that have nothing to do with his message. 8^/ -Patrick >-----Original Message----- >From: Eduardo Cavazos > >What's a functor? Can you show me an example of how they are better in some >way? > >On Mon, 14 Feb 2005 09:24:59 +0530, david rush <kumoyuki(a)gmail.com> wrote: >> > ;; In an object system like the one I made it's much shorter: >> >> Well, I used to think that way too. Believe me, you're on well-trodden >> ground here. >> >> Anyway I switched to heavy use of functors because I got tired...

3 2

Error handler in 3.0
by Eduardo Cavazos 14 Feb '05

14 Feb '05

Is there an undocumented way to specify an error handler in Gambit 3.0? I mean, instead of going into a nested REPL, can Gambit run a function that we provide? I wrote an X11 window manager with Gambit 3.0. (It's very fast! ;-) ) I run it inside Gambit in interactive mode so that I can watch the log and interact with the system if there's an error. However, if I've switched to a desktop besides the one with the xterm/REPL, and the wm crashes, I can't see the window I need... So I basically need a simple error handler that will map that window. In the meantime I've taken to running the wm from a virtual console under linux (F1, F2, F3, etc...) and switching to it whenever it crashes. For times when it's really unstable, I just use Xnest. Eduardo

2 2

Re: [gambit-list] New types
by Eduardo Cavazos 14 Feb '05

14 Feb '05

Thanks for the suggestion Marc. I think this will work perfectly. Here is the reason that I'd like to stick with the procedural representation instead of using Gambit structures directly. (define-structure point x y) (define-structure airplane pos vel) (define a (make-airplane (make-point 10 10) (make-point 0 0))) ;; Accessing the x coordinate of the airplane: (point-x (airplane-pos a)) ;; In an object system like the one I made it's much shorter: [[a pos:] x:] Getting at an object inside an object inside an object... in a struct/record system that typically accompanies Scheme implementations can be a pain in that it's so verbose. Even in C all you have to write is the name of the variable and the name of the field (plus the dot or arrow). "Different things should look different". By convention I use brackets whenever accessing or setting a field in an object. To set the value of x in the airplane above it would look like this: [[a pos:] x: 5] As opposed to this: (point-x-set! (airplane-pos a) 5) Structs in Xlib have pretty verbose names. It's nice to say this: [attr border_pixmap:] Instead of this: (XSetWindowAttributes-border_pixmap attr) I can think of more examples, but anyone that has programmed large Scheme projects surely knows this pain. In Gambit we have two styles of keywords. foo: and :foo. I use the post-colon keywords to denote fields. As another convention, I considered using pre-colon keywords to denote methods. Eduardo Cavazos On Tue, 8 Feb 2005 16:11:31 -0500, Marc Feeley <feeley(a)iro.umontreal.ca> wrote: > > Is there a way to create new primitive types in Gambit? I'm using 3.0 > > but I'm interested in answers regarding either version. > > In 3.0 and 4.0 you can use define-structure: > > Gambit Version 3.0 > > > (define-structure foo a b) > > (define x (make-foo 11 22)) > > x > #s(foo (a 11) (b 22)) > > (foo? x) > #t > > (foo? car) > #f > > (vector? x) > #f > > If you really want to keep your procedural representation, you can > access a procedure's "code" pointer to distinguish closures created > from one lambda from closures created from another lambda. Here's some > sample code: > > (define eq-procedure-code? ; this handles compiled and interpreted code > (lambda (proc1 proc2) > (if (##closure? proc1) > (and (##closure? proc2) > (if (##interp-procedure? proc1) > (and (##interp-procedure? proc2) > (eq? (##interp-procedure-code proc1) > (##interp-procedure-code proc2))) > (eq? (##closure-code proc1) > (##closure-code proc2)))) > (eq? proc1 proc2)))) > > (define make-adder > (lambda (x) > (lambda (y) > (+ x y)))) > > (define make-multiplier > (lambda (x) > (lambda (y) > (* x y)))) > > (define f (make-adder 1)) > (define g (make-adder 2)) > (define h (make-multiplier 3)) > > (pp (eq-procedure-code? f car)) ; => #f > (pp (eq-procedure-code? f g)) ; => #t > (pp (eq-procedure-code? g h)) ; => #f > > > Marc Feeley >

2 2

Global environment
by Eduardo Cavazos 11 Feb '05

11 Feb '05

Is there a way to access the global environment in Gambit 3.0? Something like a function that returns a table mapping symbols to their current values?

1 0

RE: [gambit-list] gambit 4 beta 12 on Win32
by Dan Hristodorescu 09 Feb '05

09 Feb '05

Hi Patrick, I had success with mingw (using msys shell) and I think it's a better option than cygwin since it doesn't depend on any other dlls. There are some tricks to make it work with the windows command prompt since after installation it assumes it's a unix environment and tries to execute a shell script. You'll have to create a batch file for that in order to compile with gsc. I could compile standalone exes, libraries to load in the interpreter including Meroon object system. I compile with gsc from the command prompt (the interpreter and compiler have some display problems with msys) and afterwards I use bash to compile the generated C files with gcc. After a lot of work (trying different ways to do it) I finally got a standalone exe with Meroon included and some test classes. If you need any help to set it up let me know. It does also work with Visual Studio 6 but you'll have to update the projects in the misc directory (they are old). I cannot comment on performance of the two different compilers, I only got some "line too long" errors in VS6 but it works. My choice was primarily based on the fact that mingw is open source and a lot of C libraries come with autoconf instead of VS projects. Regards, Dan Hristodorescu. -----Original Message----- I'd like to run beta 12 on Win32. Is there a technical reason a system has not been built yet? What would be the best way to build beta 12 for Win32? - Using an IDE config included in the distribution? - Using one of the Win32 shell scripts - Perhaps under Cygwin? I assume gcc is the preferable compiler, so perhaps the best choice is DJGPP and perhaps Cygwin second choice. Has anyone tried Cygwin? A google was inconclusive. Thanks, Patrick

1 0

Gambit is great: the manual should advertise better
by Bill Richter 08 Feb '05

08 Feb '05

I'm a very satisfied Gambit user, as my program became fast enough to refute some C calculations my advisor made 20 years ago. My advisor was clever enough to prove his answers were wrong, even though his program no longer runs. Thanks to Marc for offering to help optimize my code: that got me to look at my code myself and make an obvious optimization. Guillaume's profiler certainly helped locate a hotspot. But I think Gambit is a lot better than the manual would indicate to a non-expert schemer like me. I have 4 "advertising" tips: ************ blinding speed ************ The first node does some great advertising: With appropriate declarations in the source code the executable programs generated by the compiler run roughly as fast as equivalent C programs. I believe this now, with my program's speed. But it would be even better if you cross-ref'ed the node "Miscellaneous extensions", where `declarations' is explained, and if that node included Guillaume's advice to me: stick something like this at the top of your program: (declare (standard-bindings) (fixnum) (not safe) (inline) (inlining-limit 1000) (block)) One thing great about "fast as equivalent C programs" is it's the first thing a reader will see! If you bury great stuff deep in the manual, then maybe you'll restrict your users to expert programmers. The manual looks very accurate to me, but it's very technical. ************ great debugger ************ I've often felt that bad or nonexistent debuggers is the Achilles heel of Scheme. But it looks to me like Gambit has a great debugger! I think the first node "The Gambit-C system" shoul say something like: `Gambit has a good debugger. See the examples in the nodes "Debugging commands" and "Procedures related to debugging" for examining variables in case of an error, and breakpoints and single-stepping.' ************ Case sensitive ************ The node "Runtime options for all programs" says that Gambit is case-sensitive by default. I think that's a huge plus, and should be mentioned in the 1st node. I think all mathematically-inclined folks will agree with me, as one typically writes, "Let X be a set, and let x be an element of X", which turns naturally into Scheme (let ([x (car X)]) ...) I think you should target folks who use DrScheme, written by your R6RS co-Editor Matthew Flatt. DrScheme is really slow, but it has some real pedagogical advantages, I think, including case-sensitivity. I would've used Gambit years ago if I'd known it was case-sensitive. ************ read-all: don't compile data ************ I finally understood some advice Guillaume posted: don't load/include big data files: it slows down gsc/gcc unbearably. You don't need to load/include, because of Gambit extensions like read-all. This helps with my Curtis-algorithm program, but it was crucial for an example program, where I gave an independent refutation of an error my advisor made in t = 71, where I differentiated a polynomial with 75,000+ terms of length 10! In 16.6 minutes! Look at this short program: (declare (standard-bindings) (fixnum) (not safe) (inline) (inlining-limit 1000) (block)) (include "adem.scm") (define inport (open-input-file "tag-data-t-71")) (define X (read-all inport)) (close-input-port inport) (pretty-print (length X)) (pretty-print (Poly-first X)) (time (pretty-print (Poly-first (D X)))) 75299 (4 8 4 5 3 5 9 3 5 7 7) (3 4 4 2 3 5 3 5 9 7 7 7) (time (pretty-print (Poly-first (D X)))) 998093 ms real time, 16.6 minutes 997940 ms cpu time (997840 user, 100 system) 3204 collections accounting for 365525 ms real time (365690 user, 30 system) 112648202368 bytes allocated 13318 minor faults 3 major faults BTW can someone explain this figure, 112648202368 bytes allocated? Did I really use 112.6 gB? My machine "only" has 1 gB of memory. An important extension to Scheme like `read-all' maybe should be more prominently mentioned, and it's usefulness stressed. The description is fine, but I didn't find these examples very illuminating: - procedure: read-all [PORT [READER]] This procedure repeatedly calls the procedure READER with PORT as the sole argument and accumulates a list of each value returned up to the end-of-file object. The procedure `read-all' returns the accumulated list without the end-of-file object. If it is not specified, PORT defaults to the current input-port. If it is not specified, READER defaults to the procedure `read'. For example: > (call-with-input-string "3,2,1\ngo!" read-all) (3 ,2 ,1 go!) > (call-with-input-string "3,2,1\ngo!" (lambda (p) (read-all p read-char))) (#\3 #\, #\2 #\, #\1 #\newline #\g #\o #\!) > (call-with-input-string "3,2,1\ngo!" (lambda (p) (read-all p read-line))) ("3,2,1" "go!")

3 9

[gambit-list] subvector, vector-copy, vector-append need to be documented
by Bradley Lucier 08 Feb '05

08 Feb '05

Marc: You need to document the following nonstandard procedures (found by Maxim Naumov) in the Gambit documentation: vector-copy subvector vector-append These are documented for all the homogenous vector types but not for heterogeneous vectors. Brad

2 1

New types
by Eduardo Cavazos 08 Feb '05

08 Feb '05

Is there a way to create new primitive types in Gambit? I'm using 3.0 but I'm interested in answers regarding either version. My simple object system builds 'objects' out of procedures with local state. E.g.: (let ((x 10) (y 20)) (lambda args ...)) The (classic?) problem here is that my 'objects' don't look different from procedures. If I know that I'm dealing with one of my objects, I can ask it what it it is. But if I'm dealing with something that can be any Scheme type, I can ask if it's a procedure, but then I can't tell if it's one of my objects or a proc... I need a way to have a 'new' procedure type. A type of primitive item that acts just like a procedure but answers #f to the 'procedure?' predicate. I guess I am abusing procedures with this style of object system. But I'm hacking so it's OK. ;-) Ed

3 2

gambit 4 beta 12 on Win32
by Logan, Patrick D 07 Feb '05

07 Feb '05

I'd like to run beta 12 on Win32. Is there a technical reason a system has not been built yet? What would be the best way to build beta 12 for Win32? - Using an IDE config included in the distribution? - Using one of the Win32 shell scripts - Perhaps under Cygwin? I assume gcc is the preferable compiler, so perhaps the best choice is DJGPP and perhaps Cygwin second choice. Has anyone tried Cygwin? A google was inconclusive. Thanks, Patrick

2 1