December 2008 - Gambit-list - mailman.iro.umontreal.ca

RAPPEL: Soutenance de doctorat de Etienne Bergeron a 14h00
by Marc Feeley 10 Dec '08

10 Dec '08

Je vous rappelle la soutenance de la thèse de doctorat d'Etienne Bergeron à 14h00 aujourd'hui. Sujet : Compilation efficace pour FPGA reconfigurable dynamiquement Le candidat vise à démontrer la faisabilité, pour un circuit configurable de type FPGA, de se reconfigurer lui-même dynamiquement à la manière des compilateurs JIT pour augmenter son efficacité à exécuter un algorithme. Parmi les résultats de recherche, le candidat montrera comment il est parvenu à retrouver le format des fichiers de configuration d'un FPGA et comment il est capable de programmer un circuit en quelques millisecondes alors que les outils commerciaux demandent en général des minutes, voire des heures. Bienvenue à tous ! Président-Rapporteur Poulin, Pierre Directeur de Recherche David, Jean Pierre Codirecteur Feeley, Marc Membre du Jury Monnier, Stefan Examinateur Externe Thibeault, Claude, Département de génie électrique, École de Technologie Supérieure Représentante du Doyen : Bourlioux, Anne, Faculté des arts et des sciences, Département de Mathématiques et de Statistique Date MERCREDI 10 décembre, 2008 Heure 14:00 Endroit Local 3195 Pavillon André Aisenstadt 2920 Chemin de la Tour

1 0

Doctorat de Etienne Bergeron : compilation efficace pour FPGA reconfigurable dynamiquement
by JP David 08 Dec '08

08 Dec '08

Bonjour, Etienne Bergeron soutiendra sa thèse de doctorat ce mercredi à 14H00. Sujet : Compilation efficace pour FPGA reconfigurable dynamiquement Le candidat vise à démontrer la faisabilité, pour un circuit configurable de type FPGA, de se reconfigurer lui-même dynamiquement à la manière des compilateurs JIT pour augmenter son efficacité à exécuter un algorithme. Parmi les résultats de recherche, le candidat montrera comment il est parvenu à retrouver le format des fichiers de configuration d'un FPGA et comment il est capable de programmer un circuit en quelques millisecondes alors que les outils commerciaux demandent en général des minutes, voire des heures. Bienvenue à tous ! Président-Rapporteur Poulin, Pierre Directeur de Recherche David, Jean Pierre Codirecteur Feeley, Marc Membre du Jury Monnier, Stefan Examinateur Externe Thibeault, Claude, Département de génie électrique, École de Technologie Supérieure Représentante du Doyen : Bourlioux, Anne, Faculté des arts et des sciences, Département de Mathématiques et de Statistique Date MERCREDI 10 décembre, 2008 Heure 14:00 Endroit Local 3195 Pavillon André Aisenstadt 2920 Chemin de la Tour -- -------------------------------------------------- Ph.D. Jean Pierre David Département de Génie Électrique École Polytechnique de Montréal http://www.grm.polymtl.ca/~david Tél. (1)514-340-4711 p2009

1 0

Wiki links for documentation
by Marc Feeley 05 Dec '08

05 Dec '08

I have created a wiki namespace for the documentation and will use the following convention for links: [[Documentation:procedure getenv|procedure <tt>getenv</tt>]] [[Documentation:special form time|special form <tt>time</tt>]] That way the name of the procedure or special form in the link will be in tt font. I don't know if it is possible for the name to be in tt font in the page title, but I could not find how to do it. Moreover, it seems that names containing < and > and perhaps other special characters like the vertical bar can't be used in a link (and thus can't appear in a page title). I have tried [[Documentation:procedure <]] But that doesn't work (the system thinks that this is not a wiki link). Anyone here know if there is a solution to this problem? Marc

1 1

Your changes to treatment of (begin)
by Bradley Lucier 05 Dec '08

05 Dec '08

An empty (begin) is not valid r5rs syntax, yet it's now accepted by both the interpreter and the compiler. I don't like language extensions like this that offer little in the way of utility but will screw up people porting code from gambit to other implementations, in the same way that PLT's mapping of () to '() screwed up people moving from PLT to gambit (I don't know whether PLT still does that). Brad

2 2

const qualifier
by Marco Benelli 04 Dec '08

04 Dec '08

Hi all. It is possible to specify a "const" qualifier in c-define-type ? I know it's not in ANSI standard, but it seem quite common in gcc-based projects. Thanks in advance. --marco

2 1

Re: [gambit-list] Adding docs to wiki?
by Marc Feeley 02 Dec '08

02 Dec '08

I looked into what was necessary to write a Texinfo plugin for MediaWiki (which is used by the Gambit wiki). It turns out to be simpler than I thought. So I wrote a plugin which handles <texinfo> ...Texinfo document... </texinfo> in any wiki page. The plugin calls up texi2html to convert the Texinfo document to HTML (currently everytime the page is accessed). Here are 3 pages written with the <texinfo> tag: http://dynamo.iro.umontreal.ca/~gambit/wiki/index.php/Procedure_getenv http://dynamo.iro.umontreal.ca/~gambit/wiki/index.php/Procedure_current-time http://dynamo.iro.umontreal.ca/~gambit/wiki/index.php/Special_form_time Go to one of these pages and click on "edit" to see what the source document looks like. The Texinfo document is wrapped inside <texinfo> and </texinfo> tags. I'd like to get some feedback on these pages before I put the rest of the Gambit manual on the wiki. Some issues: 1) What name should the pages have? I use the Procedure_ and Special_form_ prefixes so that a name can easily be found. But perhaps it is better to have a unique prefix (Documentation_of_ or simply nothing, but this might cause a name clash) so that the page can be accessed directly without knowing if the name refers to a procedure or a special form. This would make the wiki documentation easy to access from an IDE. I can easily imagine the IDE adding to the error messages a hyperlink to the wiki documentation of the procedure which raised the exception. 2) Where should the documentation which is outside the Texinfo @deffn forms be put? For example, general discussion of I/O, or debugging, or the compiler, general index, concept index, etc. 3) How should hyperlinks work within the wiki? 4) Can the HTML's prettiness be improved? 5) How can the wiki documentation be copied back to the Gambit manual? When should this happen? Marc On 30-Nov-08, at 2:31 PM, Ali wrote: > 2008/11/30 Marc Feeley <feeley(a)iro.umontreal.ca> > Having the documentation on the wiki or some user editable place > would be useful. However, I think the documentation should be > written in one "universal" markup language so that all forms of the > documentation can be generated from it (.pdf, .html, .info, and > "wiki" editable). Currently that markup language is texinfo. > > Yeah making a copy of the docs is clearly a bad idea looking back at > it. The impression I got from this thread is that it would be a nice > thing to have some website written using Gambit which had 2 basic > functions: > 1) Convert texinfo to some suitable html / wiki format and cache it. > 2) Allow users to edit the html / wiki format, which then gets > converted back to texinfo format and committed to the docs and > regenerated again using (1). > > The trick here would be making sure the texinfo -> wiki -> texinfo > conversion cycle did not change texinfo to something different, but > hopefully this could be done using html attributes. > > I didn't "bagsie" this myself because I don't know whether this is > beyond me and I can't put an eta on learning all of the stuff behind > it, but I'd like to have a go at it some time if no one has by that > point. > > Moreover, eventually I would like the documentation for the > procedures and special forms to be inside the source code and the > examples in the documentation should be tested for consistency when > the regression tests are run. I'm unsure how all of this can work. > Is there a documentation maintenance system that supports all of > this? Otherwise can one be built? > > Marc > > Similar to JavaDoc? I like the idea, and I'm optimistic that the > tools could be written, but this involves questions of file size and > what type of docs would be in the source, so I'm not really inclined > to comment! > > ps. Playing around with Gambit a bit more now has only made me > happier, I've been very pleasantly surprised by some things. Thanks > for the good work, keep it up! >

4 4

location of libgambcgsi.so
by Marijn Schouten (hkBst) 01 Dec '08

01 Dec '08

-----BEGIN PGP SIGNED MESSAGE----- Hash: SHA1 Hi list, which line of configure.ac or one of the makefile.in's determines what the exact path, as given by the first line of output of the command "readelf -d /usr/bin/gsi", is at which the executable gsi searches for libgambcgsi.so in the case that gambit has been configured with --enable-shared? Thanks, Marijn - -- Marijn Schouten (hkBst), Gentoo Lisp project, Gentoo ML <http://www.gentoo.org/proj/en/lisp/>, #gentoo-{lisp,ml} on FreeNode -----BEGIN PGP SIGNATURE----- Version: GnuPG v2.0.9 (GNU/Linux) Comment: Using GnuPG with Mozilla - http://enigmail.mozdev.org iEYEARECAAYFAkkvUS0ACgkQp/VmCx0OL2xLvACgsfu1/jL9knChwRm6Mm+0sDk/ f0oAoIR9xtjtgiZNiXB4BOBsRX3Reval =bk9I -----END PGP SIGNATURE-----

2 6

Can't use make update
by Bradley Lucier 01 Dec '08

01 Dec '08

This doesn't work (from history): > 7 23:43 tar zxf gambc-v4_3_1-devel.tgz > 9 23:43 cd gambc-v4_3_1-devel > 10 23:43 ./configure CC='gcc -mcpu=970 -m64' --enable-single- > host > 11 23:44 make -j 2 bootstrap > 12 23:48 make update It fails with <reconfiguring here> > making all in include > cd .. && CONFIG_FILES=include/makefile CONFIG_HEADERS= ./config.status > config.status: creating include/makefile > major=`echo v4.3.2 | sed -e"s&v$[^.]*$\.$[^.]*$\.$[^.]*$.*& > \1&"`; \ > minor=`echo v4.3.2 | sed -e"s&v$[^.]*$\.$[^.]*$\.$[^.]* > $.*&\2&"`; > \ > revision=`echo v4.3.2 | sed -e"s&v$[^.]*$\.$[^.]*$\.\ > ([^.]*\).*&\3&" > `; \ > version_num=`eval expr "$ 100000 \\* $major $ + $ 1000 \ > \* $minor $ > + $revision"`; \ > echo "#error \"The version indicated in ___VERSION is not > supported by gambit.h\"" > gambit-not$version_num.h; \ > echo "#error \"You are probably compiling C code generated > by a Gambit compiler earlier or later than v4.3.2\"" >> gambit-not > $version_num.h; \ > echo "#error \"___VERSION will be set to $version_num to > reduce other errors\"" >> gambit-not$version_num.h; \ > echo "#undef ___VERSION" >> gambit-not$version_num.h; \ > echo "#define ___VERSION $version_num" >> gambit-not > $version_num.h; \ > echo "#include \"gambit.h\"" >> gambit-not$version_num.h > cd .. && CONFIG_FILES=include/gambit.h CONFIG_HEADERS= ./config.status > config.status: creating include/gambit.h > making all in lib > cd .. && CONFIG_FILES=lib/makefile CONFIG_HEADERS= ./config.status > config.status: creating lib/makefile > ../gsc-comp -:=.. -f -c -check _io.scm > gcc -mcpu=970 -m64 -I../include -I. -no-cpp-precomp -Wall -W -Wno- > unused -O1 -fno-math-errno -fschedule-insns2 -fno-trapping-math - > fno-strict-aliasing -fwrapv -fomit-frame-pointer -fPIC -fno-common - > DHAVE_CONFIG_H -D___PRIMAL -D___LIBRARY -D___GAMBCDIR="\"/usr/local/ > Gambit-C/v4.3.2\"" -D___SYS_TYPE_CPU="\"powerpc\"" - > D___SYS_TYPE_VENDOR="\"apple\"" -D___SYS_TYPE_OS="\"darwin9.5.0\"" - > c _io.c > In file included from ../include/gambit.h:12, > from _io.c:1278: > ../include/gambit-not403002.h:1:2: error: #error "The version > indicated in ___VERSION is not supported by gambit.h" > ../include/gambit-not403002.h:2:2: error: #error "You are probably > compiling C code generated by a Gambit compiler earlier or later > than v4.3.2" > ../include/gambit-not403002.h:3:2: error: #error "___VERSION will > be set to 403002 to reduce other errors" > make[2]: *** [_io.o] Error 1 > make[1]: *** [all-recursive] Error 1 > make: *** [update] Error 2

1 0