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

profiling code being interpreted by gsi
by lowly coder 31 Mai '09

31 Mai '09

Hi! I have a largeish gambit application. It's interpreted under gsi, and is taking up around 25% of my dual core laptop. I would like to profile to see what functions are taking up all the time. How can I do this? (Searching for profiling in the gambit documentation brings up stuff focused on how to profile scheme code that has been compiled to C, but not on profiling of interpreted code.) Thanks!

4 4

[Fwd: Measuring FSF gcc from 4.1.2 to today on various benchmarks.]
by Bradley Lucier 29 Mai '09

29 Mai '09

I've run some statistics with various versions of gcc and sent the following message to the gcc mail list, which is archived at http://gcc.gnu.org/ml/gcc/ Perhaps some of you may be interested. Brad -------- Forwarded Message -------- > From: Bradley Lucier <lucier(a)math.purdue.edu> > To: gcc(a)gcc.gnu.org > Cc: Bradley Lucier <lucier(a)math.purdue.edu> > Subject: Measuring FSF gcc from 4.1.2 to today on various benchmarks. > Date: Fri, 29 May 2009 15:58:00 -0400 > > I've put at > > http://www.math.purdue.edu/~lucier/bugzilla/9/ > > some compile-time and run-time statistics related to PR 39157 and PR > 33928 and compile times and run times for the programs in the Gambit > Scheme benchmark suite. The statistics are for 4.1.2 release, 4.2.4 > release, 4.3.3 release, 4.4.1 20090522, 4.5.0 20090521 (revision > 147758), and 4.5.0 20090521 (revision 147758) with > -fno-forward-propagate; they use (mainly) the set of options > > -O1 -fno-math-errno -fschedule-insns2 -fno-trapping-math -fno-strict-aliasing -fwrapv -fomit-frame-pointer -fPIC -fno-common -mieee-fp > > on a Core 2 quad processor (running basically nothing else at the time). > > I would conclude from the statistics that, right now, the cost of > including -fforward-propagate in -O1 overrides any performance benefit > that may result. > > Brad >

1 0

Gambit positively spotted on slashdot
by Adrien Piérard 29 Mai '09

29 Mai '09

Recently, my friend John wrote a webserver in CL that made its way to slashdot. That's a good thing for FP and lisp based languages. The article can be seen here: http://tech.slashdot.org/article.pl?sid=09/05/25/1553220 Moreover, in the comments, someone inserted Gambit in the discussion, using infix notation: http://tech.slashdot.org/comments.pl?sid=1244601&cid=28087817 I believe this too is a very good intervention in favor of Scheme and particularly Gambit. P! -- Français, English, 日本語, 한국어

3 2

Question on dynamic-wind
by Jeremie Lasalle Ratelle 29 Mai '09

29 Mai '09

I am trying to use dynamic-wind to execute code on thread suspension and thread restore but it doesn't seem to work. Is it possible or is there a way that I can do this? I need it to call a thunk just before it will call its continuation or just after it has been suspended. What is the correct way to do this?

1 0

define-type
by lowly coder 28 Mai '09

28 Mai '09

Hey, Where is define-type documented? I can't find it in r5rs or in the gambit documentation (though it is briefly mentioned in a tour of scheme through gambit). define-type is really cool and I'm starting to run against my limited knowledge of it. Thanks!

3 8

Viewing CPS expansion?
by James Long 25 Mai '09

25 Mai '09

`gsc' supports -expansion, but it seems to show the basic expansion into normalized gambit code. Is there any way to show the fully CPS-transformed forms of the code? Purely for research purposes. - James

6 17

Closures, continuations, strategies, and actual measurements
by D.McClain 24 Mai '09

24 Mai '09

There is a very revealing paper by Clinger, Hartheimer, and Ost, "Implementations Strategies for First Class Continuations", that provides detailed examinations of various strategies, the code generated to support them, and actual measurements showing the effects of each implementation variation. http://www.springerlink.com/content/h5808n962434j275/fulltext.pdf This paper, along with Marc's recent explanation about the cost of CPS, should go a long way toward explaining why CPS code may be slower than direct form code. Dr. David McClain Sr. VP, Embedded Systems Asyrmatos Inc. Boston & Tucson phone: 520-529-2437 cell: 520-390-3995 web: www.asyrmatos.com e-mail: dbm(a)asyrmatos.com

3 5

for each, hash table, order
by lowly coder 23 Mai '09

23 Mai '09

i realize that order is fairly meaningless in the context of a hashtable however, if i for-each twice through a hashtable, and there are no writes (add or remove) to the hashtable in between them, am I, by the way that gambit is implemented, guaranteed to hit the elemements in the same order for the two times? thanks!

2 1

READ for things other than lists
by FFT 22 Mai '09

22 Mai '09

It appears that while WRITE works for just about anything, READ does not work for vectors, structures and probably other things. Is there a simple way to convince READ to recognize what WRITE has written? (Clojure does it)

2 1

FFI
by FFT 22 Mai '09

22 Mai '09

I asked this in comp.lang.scheme already: Is it possible to declare C types that are structs that include arrays (not pointers)? Also, what about bit fields? The answer I got in comp.lang.scheme is that one needs to use lists of strings (?!) for the former. Perhaps I'm missing something, but this seems to be very kludgy, even if it does work. Structs that have arrays are pretty common in C APIs (Bit fields happen too) What's the best approach here?

2 5