Gambit-list Octobre 2008

gambit-list@iro.umontreal.ca

30 participants
36 discussions

Edge triggered I/O with libevent
by Adam Langley 04 Avr '13

04 Avr '13

Gambit uses the select call on POSIX systems to multiplex IO operations. This is fine for small numbers of file descriptors, but falls apart quickly with large numbers[1]. libevent[2] is a wrapper library around a number of modern IO multiplexing systems (epoll on Linux, kqueue on the BSDs as well as poll and select). Attached is a patch which converts Gambit to using edge triggered I/O with epoll and libevent. This also requires a patch (attached) of libevent since the standard distribution doesn't include edge triggering support and the upstream appears to be on holiday. (epoll is required because I don't have a BSD box on which to develop edge triggering support for kqueue) This patch (for Gambit) preprocesses away to nothing unless --enable-libevent is passed to the configure script. (Note that, since this patches the configure.ac script, you'll need to run autoconf to regenerate). I hope I've not broken anything, but at this stage I've only tested that the console and tcp ports work. This patch (in the future) should also allow people, who write programs which use socket calls directly, to integrate into the main event loop at some point. Testing is more than welcome and, if you find a form of IO which is broken, please email me. [1] http://monkey.org/~provos/libevent/libevent-benchmark2.jpg [2] http://monkey.org/~provos/libevent/ -- Adam Langley agl(a)imperialviolet.org http://www.imperialviolet.org (+44) (0)7906 332512 PGP: 9113 256A CC0F 71A6 4C84 5087 CDA5 52DF 2CB6 3D60

2 2

Interactive 3D graphics with Gambit-C
by Alex Sandro Queiroz e Silva 06 Jan '10

06 Jan '10

Hallo, I am thinking about doing some basic 3D graphics with Gambit-C. Although I am sure code generated by Gambit-C is fast enough[1], I am a bit worried about the garbage collector. What's the type of garbage collector used by Gambit? If it kicks in during interactive rendering, should I expect annoying pauses? [1] - http://www.ventonegro.org/2008/09/the-need-for-speed/ -alex http://www.ventonegro.org/

5 7

I/O puzzle...
by Marc Feeley 09 Jui '09

09 Jui '09

Puzzle... what's the bug in the following code? See end of message for answer. (define http-server-start! (lambda (hs stop?) (let ((server-port (open-tcp-server (list port-number: (server-port-number hs) backlog: 128 reuse-address: #t server-address: '#u8(127 0 0 1) ; on localhost interface only char-encoding: 'ISO-8859-1)))) (input-port-timeout-set! server-port 3) (accept-connections hs server-port stop?)))) (define accept-connections (lambda (hs server-port stop?) (let loop () (let ((connection (read server-port))) (if (eof-object? connection) (if (stop?) (close-input-port server-port) (loop)) (begin (if (server-threaded? hs) (let ((dummy-port (open-dummy))) (parameterize ((current-input-port dummy-port) (current-output-port dummy-port)) (thread-start! (make-thread (lambda () (serve-connection hs connection)))))) (serve-connection hs connection)) (loop))))))) . . . . . . . . Answer: the (input-port-timeout-set! server-port 3) is at the wrong place. It sets a timeout ***3 seconds in the future***. So after 3 seconds have passed all calls to read in accept-connections will not block (and lead to a tight, CPU intensive loop). The correct approach is to put the call to (input-port-timeout-set! server-port 3) just before the call to read. Puzzle #2: why did Gambit chose to define timeouts that are not relative to the moment that an I/O function is called? Marc

2 3

Relaxing moment
by Guillaume Cartier 17 Nov '08

17 Nov '08

For a relaxing moment using Gambit, try evaluating the following : (inexact->exact 2.2) ---------- the 2476979795053773/1125899906842624 result shows that Gambit does indeed smoke some powerfull stuff!

4 4

lib/sys.scm seems to be unused holdover
by Christian Jaeger 31 Oct '08

31 Oct '08

It seems lib/sys.scm can be removed? At least Gambit builds and works without it. Christian.

1 0

Table iterators
by David St-Hilaire 31 Oct '08

31 Oct '08

Hi! I just found out that their are the ##table-foldl and ##table-for-each internal procedure that are actually essential for making a program that use table efficient, but why are these only internals, and not external functions? I would have really liked to know about these before now! hehe Thanks David

3 2

(case)
by Christoph Bauer 31 Oct '08

31 Oct '08

thanks for all the responses in the string->number thread. I have another question: what is the correct result of (case ".ab" ((".ab") 1) (else 2)) Indeed gsi gives another answer as scheme48. (maybe eq? vs equal?) Christoph

5 5

Patch to improve elementary functions for complex and exact arguments
by Bradley Lucier 31 Oct '08

31 Oct '08

Marc: Here is a diff containing the following changes to _num.scm: new macros: macro-exp-1/2, macro-exp--1/2, macro-log-2 new function: ##flonum.full-precision, which checks that (abs x) is finite and above the largest subnormal number. In ##log: Define exact-log, which gives a finite answer for extremely large and extremely small exact arguments, and better relative accuracy for exact rationals near 1. It is slower than the old behavior of log on exact rationals near 1, so if you just want a quick and dirty approximation, use (log (exact->inexact x)). Define complex-log-magnitude, which computes (log (magnitude x)) carefully in one step instead of in two steps. in ##atan2: Return exact 0 if x and y are exact, x is positive, and y is 0. Carefully scale arguments before handling off to ##flonum.atan so more answers are finite. Remove some checks, put them in calling functions. in atan: add checks before calling ##atan and ##atan2. in ##sqrt: Define complex-sqrt-magnitude, which carefully calculates (sqrt (magnitude x)) in one step instead of two. For complex arguments, return an exact result where possible, e.g., (##sqrt +2i) => 1+i. I've included a test program below, and the results of that test program in old-results and new-results. Many previous +nan.0 and +inf. 0 results are now finite; finite results that differ in more than the last bit or so have more accuracy. Brad

2 6

Re: [gambit-list] Gambit Cygwin package
by Nathan Thern 31 Oct '08

31 Oct '08

On Thu, Oct 30, 2008 at 8:24 PM, Mikael More <X@X> wrote: > Hi! > Thank you for porting Gambit into Cygwin! http://www.liquid.spod.org/~nthern/cygwin/ > Although, on my machine, I didn't get it running, neither the package > (compile-file complains it can't start external process) nor the Gambit > sources (compile-file and shell-command fire cmd.exe instead of /bin/sh, > even though i updated COMSPEC and SHELL environment variables). > I just wanted to check with you if you know how to get it going. > Thanks! > > Mikael Interesting. All my testing so far consisted of calling "gsc -link foo.scm" to create c files and then compiling them with "gcc ...". Plain old "gsc foo.scm" does indeed exhibit the behavior you described. Are you able to run "gsc -link foo.scm"? If so, then I would say you at least "got it running". Not sure where the problem lies. Since the cygwin system call API is POSIX compliant, I'm going to guess that it's a configuration script problem. That is, I'm guessing that somewhere configure assumes something like "if we have windows.h then without exception make all system calls through cmd.exe". I'm looking in the source but haven't found it yet.

1 0

Experimental tree-shaker
by Marc Feeley 31 Oct '08

31 Oct '08

I have implemented a simple quick-and-dirty tree-shaker for those interested in experimenting with such a thing. To try it out, update your Gambit to the latest sources. The tree- shaker is disabled by default. To use it you must compile your program this way: % gsc -e '(set! c#targ-tree-shake? #t)' test.scm The tree-shaker will only remove from a compiled file the procedures which are not reachable *and* defined in the scope of a (declare (block)). So if you compile this program: (declare (block)) (define (f x) (+ (* x x) (- x 1))) (define (g y) (f (f y))) (pp 123) both f and g are eliminated. If the (pp 123) is changed to (pp (g 123)) then there is code generated for f, but ***not for g*** because the compiler has inlined g so that the main call is really (pp (f (f 123))). For the experimenter... implement a Scheme library in lib.scm and compile your program like so: (declare (block)) (include "lib.scm") ...your program here... The last step is to link with a minimal Gambit runtime library containing only _kernel.scm . Marc

1 0

2024

2023

2022

2021

2020

2019

2018

2017

2016

2015

2014

2013

2012

2011

2010

2009

2008

2007

2006

2005

2004

Gambit-list Octobre 2008