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

porting to system without tty or standard IO
by Scott Graham 29 Mai '05

29 Mai '05

Hi Has anyone tried porting Gambit to a system without a tty and/or without the low-level file descriptor level? I'm trying to port to an embedded system that's quite similar to USE_WIN32 as far as os_*.[ch] are concerned, however, it has neither a TTY, nor stdin. I had a try, and got as far as gsi compiling, but I'm getting an exception that I don't understand early on startup. So, just wondering if anyone has tried this, or has suggestions on plan of attack. Ideally, I'd like to have the system work over a socket connected to the host PC rather than expecting a variety of streams to be available "locally". thanks, scott

2 2

64-bit bignum times
by Bradley Lucier 25 Mai '05

25 Mai '05

Marc: I've run my bignum tests on a 2GHz G5 running MacOS 10.4.1 with Apple's compiler gcc version 4.0.0 (Apple Computer, Inc. build 5018) from CVS sources, and on a 2.4GHz Opteron running the latest Red Hat Enterprise Linux with gcc-4.0.0. On the Mac, Gambit was configured with env CC='gcc -m64 -mcpu=970 -force_cpusubtype_ALL' ./configure -- enable-single-host and on the Opteron, it was just ./configure --enable-single-host but -m64 and -mcpu=whatever are the defaults. The source and the result files are below. Basically, cycle for cycle, the Mac beats the Opteron; it's also available at higher speeds and is cheaper. A lot of the time is spent in FFTs for the bignum multiplies, so this is not a bad test for 64-bit scientific computing capabilities. Brad PS: I Bcc'ed the other people receiving this so their e-mail addresses won't show up on the public mail list.

1 0

running remote jobs with `time'
by Bill Richter 22 Mai '05

22 Mai '05

I haven't been able to run niced background jobs, and I think I finally know why, because `time' wants to print to the console which no longer exists. But I'm not quite there yet: The Gambit manual says `time' uses the interaction channel: - special form: time EXPR The `time' special form evaluates EXPR and returns the result. As a side effect it displays a message on the interaction channel which indicates how long the evaluation took (in real time and cpu time), how much time was spent in the garbage collector, how much memory was allocated during the evaluation and how many minor and major page faults occured (0 is reported if not running under UNIX). Now the node "Runtime options" says I can change the interaction channel: Both `gsi' and `gsc' as well as executable programs compiled and linked using `gsc' take a `-:' option which supplies parameters to the runtime system. `-' The REPL interaction channel will be standard input and standard output. I have a 1-line file "6.scm" (time (pretty-print 6)) So to me, the manual says that % gsi -:- 6.scm > six-out should print the time info as well as the output 6 to the file "six-out". But it doesn't: (banach)PolyTree> gsi -:- 6.scm > six (time (pretty-print 6)) 0 ms real time 0 ms cpu time (0 user, 0 system) no collections 104 bytes allocated no minor faults 5 major faults (banach)PolyTree> more six-out 6 Now it doesn't make much difference where the time info goes, unless I'm running a job overnight, nice-ed in the bg. These nice-ed bg jobs of mine hang, run for days, and I think because they want to print to the console, which no longer exists. So people must know how to solve this problem. Help!

4 6

64-bit compile on MacOS Tiger
by Bradley Lucier 21 Mai '05

21 Mai '05

Marc: I installed Tiger 10.4.1 and built Gambit-C with CC='gcc -m64'. Everything seemed to work, but I got the following warnings: gcc -m64 -I../include -I. -no-cpp-precomp -Wall -W -Wno-unused -O1 - fno-math-errno -fschedule-insns2 -fno-trapping-math -fno-strict- aliasing -fomit-frame-pointer -fPIC -fno-common -DHAVE_CONFIG_H - D___PRIMAL -D___LIBRARY -D___GAMBCDIR=\"/usr/local/Gambit-C\" -c os_dyn.c os_dyn.c: In function 'dynamic_load_module': os_dyn.c:258: warning: 'NSIsSymbolNameDefined' is deprecated (declared at /usr/include/mach-o/dyld.h:150) os_dyn.c:260: warning: 'NSLookupAndBindSymbol' is deprecated (declared at /usr/include/mach-o/dyld.h:158) os_dyn.c: In function '___make_c_closure': os_dyn.c:1286: warning: cast from pointer to integer of different size os_dyn.c:1288: warning: cast from pointer to integer of different size os_dyn.c:1289: warning: cast from pointer to integer of different size os_dyn.c:1292: warning: cast from pointer to integer of different size os_dyn.c:1294: warning: cast from pointer to integer of different size os_dyn.c:1295: warning: cast from pointer to integer of different size os_dyn.c: In function '___is_a_c_closure': os_dyn.c:1428: warning: cast from pointer to integer of different size os_dyn.c:1430: warning: cast from pointer to integer of different size os_dyn.c:1431: warning: cast from pointer to integer of different size This looks like a real problem; is there any info I can give to help you out with this? Brad

1 2

Problem with compiled code using tables
by Bradley Lucier 21 Mai '05

21 Mai '05

Marc: There's a very nice little thread in comp.lang.lisp about how to calculate the number of partitions of a positive integer. The fastest algorithm by far uses Common Lisp's hash tables, so I thought I would try it with Gambit's tables, but it doesn't seem to work when compiled. Here is the code: (define (iota-reverse n) (let loop ((i 1) (result '())) (if (> i n) result (loop (+ i 1) (cons i result))))) (define (make-list n #!optional init) (vector->list (make-vector n init))) (define (break-partitions n break) (or (table-ref *break-cache* (cons n break) #f) (let ((result (cond ((= break 1) (list (make-list n 1))) ((= break n) (list (list n))) (else (let ((rem (- n break))) (map (lambda (i) (cons break (break-partitions rem i))) (iota-reverse (min break rem)))))))) (table-set! *break-cache* (cons n break) result) result))) (define (breaks n) (apply append (map (lambda (break) (break-partitions n break)) (iota-reverse n))))) It gives the correct result when interpreted, but not when compiled: [descartes:~/programs/gambc40b13] lucier% gsi loading /usr/local/Gambit-C/gambcext.scm Gambit Version 4.0 beta 13 > (load "partition.scm") "/Users/lucier/programs/gambc40b13/partition.scm" > (pp (breaks 10)) ((10) (9 (1)) (8 (2)) (8 (1 1)) (7 (3)) (7 (2 (1))) (7 (1 1 1)) (6 (4)) (6 (3 (1))) (6 (2 (2)) (2 (1 1))) (6 (1 1 1 1)) (5 (5)) (5 (4 (1))) (5 (3 (2)) (3 (1 1))) (5 (2 (2 (1))) (2 (1 1 1))) (5 (1 1 1 1 1)) (4 (4 (2)) (4 (1 1))) (4 (3 (3)) (3 (2 (1))) (3 (1 1 1))) (4 (2 (2 (2)) (2 (1 1))) (2 (1 1 1 1))) (4 (1 1 1 1 1 1)) (3 (3 (3 (1))) (3 (2 (2)) (2 (1 1))) (3 (1 1 1 1))) (3 (2 (2 (2 (1))) (2 (1 1 1))) (2 (1 1 1 1 1))) (3 (1 1 1 1 1 1 1)) (2 (2 (2 (2 (2)) (2 (1 1))) (2 (1 1 1 1))) (2 (1 1 1 1 1 1))) (2 (1 1 1 1 1 1 1 1)) (1 1 1 1 1 1 1 1 1 1)) > (define a (time (breaks 500))) (time (breaks 500)) 83976 ms real time 71372 ms cpu time (69503 user, 1869 system) 33 collections accounting for 7657 ms real time (6264 user, 124 system) 764582700 bytes allocated no minor faults no major faults [descartes:~/programs/gambc40b13] lucier% gsi loading /usr/local/Gambit-C/gambcext.scm Gambit Version 4.0 beta 13 > (load "partition") "/Users/lucier/programs/gambc40b13/partition.o2" > (pp (breaks 10)) ((10) (9 (1)) (8 (2)) (8 (1 1)) (7 (3)) (7 (2 (1))) (7 (1 1 1)) (6 (4)) (6 (3 (1))) (6 (2 (2)) (2 (1 1))) (6 (1 1 1 1)) (5 (5)) (5 (4 (1))) (5 (3 (2)) (3 (1 1))) (5 (2 (2 (1))) (2 (1 1 1))) (5 (1 1 1 1 1)) (4 (4 (2)) (4 (1 1))) (4 (3 (3)) (3 (2 (1))) (3 (1 1 1))) (4 (2 (2 (2)) (2 (1 1))) (2 (1 1 1 1))) (4 (1 1 1 1 1 1)) (3 (3 (3 (1))) (3 (2 (2)) (2 (1 1))) (3 (1 1 1 1))) (3 (2 (2 (2 (1))) (2 (1 1 1))) (2 (1 1 1 1 1))) (3 (1 1 1 1 1 1 1)) (2 (2 (2 (2 (2)) (2 (1 1))) (2 (1 1 1 1))) (2 (1 1 1 1 1 1))) (2 (1 1 1 1 1 1 1 1)) (1 1 1 1 1 1 1 1 1 1)) > (define a (time (breaks 500))) *** ERROR IN break-partitions, "partition.scm"@15.7 -- (Argument 1) Instance of #<type #2 table> expected (table-ref 0 '(26 . 16) #f) 1> ,e n = 26 break = 16 (current-exception-handler) = primordial-exception-handler (current-input-port) = '#<input-output-port #3 (console)> (current-output-port) = '#<input-output-port #3 (console)> (current-directory) = "/Users/lucier/programs/gambc40b13/" 1> ,y 0 break-partitions "partition.scm"@15.7 (table-ref *break-cache... 1> ,i #<procedure #4 break-partitions> = (lambda (n break) (or (table-ref *break-cache* (cons n break) #f) (let ((result (cond ((= break 1) (list (make-list n 1))) ((= break n) (list (list n))) (else (let ((rem (- n break))) (map (lambda (i) (cons break (break-partitions rem i))) (iota-reverse (min break rem)))))))) (table-set! *break-cache* (cons n break) result) result))) 1> (table-length *break-cache*) 1404 1> Do you know what's happening? This is in MacOS 10.4.1, Apple's gcc-4.0, default configure and build. I used the '(debug) option to compile the code (which works very nicely, by the way). Brad

1 3

Re: [gambit-list] Problem with compiled code using tables
by Bradley Lucier 20 Mai '05

20 Mai '05

On May 20, 2005, at 5:51 AM, Marc Feeley wrote: > Your code is incomplete... where is the definition of *break-cache*? (define *break-cache* (make-table))

1 0

64-bit timings for beta 12
by Bradley Lucier 18 Mai '05

18 Mai '05

1. Timing for 10,000 terms of the series for e 2. Timing for 10,000 terms of the atan 1/5 - c*atan(1/239) for pi 3. timing for 655360 decimal digits of pi, using base 10 (brent- salamin) 4. Timing for 655360 decimal digits of pi, using base 2 (brent-salamin) 64-bit timings: (time (binary-splitting-compute-e size)) 2205 ms real time 1781 ms cpu time (1673 user, 108 system) 61 collections accounting for 114 ms real time (77 user, 29 system) 617213896 bytes allocated no minor faults no major faults (time (binary-splitting-compute-pi size)) 2433 ms real time 2125 ms cpu time (2004 user, 121 system) 77 collections accounting for 133 ms real time (95 user, 32 system) 768422984 bytes allocated no minor faults no major faults (time (pi-brent-salamin)) 178434 ms real time 104868 ms cpu time (91853 user, 13015 system) 907 collections accounting for 8191 ms real time (1238 user, 3796 system) 8573150600 bytes allocated no minor faults no major faults (#<procedure #2 +> .028627000000010838) (#<procedure #3 -> .05454900000000862) (#<procedure #4 *> 20.701710000000013) (#<procedure #5 quotient> 24.720575999999948) (#<procedure #6 integer-sqrt> 59.353834) (time (pi-brent-salamin)) 168068 ms real time 82734 ms cpu time (71701 user, 11033 system) 765 collections accounting for 9195 ms real time (1106 user, 3246 system) 6625855176 bytes allocated no minor faults no major faults (#<procedure #2 +> .02847199999995098) (#<procedure #3 -> .05393800000000226) (#<procedure #4 *> 19.663693000000066) (#<procedure #5 quotient> 3.4773299999999097) (#<procedure #6 integer-sqrt> 59.50296600000014) 547506 "/Users/lucier/programs/gambc40b12/../gambc40b11/gambit-trace2.scm" 32-bit timings (time (binary-splitting-compute-e size)) 5165 ms real time 2766 ms cpu time (2624 user, 142 system) 63 collections accounting for 161 ms real time (74 user, 31 system) 607635944 bytes allocated no minor faults no major faults (time (binary-splitting-compute-pi size)) 6287 ms real time 3375 ms cpu time (3221 user, 154 system) 80 collections accounting for 228 ms real time (93 user, 34 system) 762781320 bytes allocated no minor faults no major faults (time (pi-brent-salamin)) 272716 ms real time 138967 ms cpu time (122341 user, 16626 system) 925 collections accounting for 10065 ms real time (1080 user, 4158 system) 8568010464 bytes allocated no minor faults no major faults (#<procedure #2 +> .030378000000034433) (#<procedure #3 -> .061922999999978856) (#<procedure #4 *> 27.754853999999927) (#<procedure #5 quotient> 33.627896000000014) (#<procedure #6 integer-sqrt> 77.48350500000001) (time (pi-brent-salamin)) 139892 ms real time 98029 ms cpu time (85901 user, 12128 system) 767 collections accounting for 5740 ms real time (826 user, 3016 system) 6621118600 bytes allocated no minor faults no major faults (#<procedure #2 +> .02661999999995146) (#<procedure #3 -> .05482300000002738) (#<procedure #4 *> 23.816316000000086) (#<procedure #5 quotient> 4.030742000000004) (#<procedure #6 integer-sqrt> 70.09432399999997) 547506

1 0

Struct alignment on Windows
by MichaelL＠frogware.com 16 Mai '05

16 Mai '05

The old Visual Studio make files force struct alignment to 4 bytes. Is that necessary today, or just a holdover from bygone days?

1 0

Warnings Compiling Gambit Under Windows
by MichaelL＠frogware.com 16 Mai '05

16 Mai '05

Gambit is generating some warnings under Windows. Some of them look like real errors: 1 d:\tools\Scheme\proj\gambc40b13\lib\os_tty.c(7344) : warning C4113: 'BOOL (__stdcall *)()' differs in parameter lists from 'PHANDLER_ROUTINE' 2 d:\tools\Scheme\proj\gambc40b13\lib\os_tty.c(7362) : warning C4113: 'BOOL (__stdcall *)()' differs in parameter lists from 'PHANDLER_ROUTINE' 3 d:\tools\Scheme\proj\gambc40b13\lib\os_time.c(234) : warning C4307: '*' : integral constant overflow 4 d:\tools\Scheme\proj\gambc40b13\lib\os_time.c(234) : warning C4244: 'initializing' : conversion from 'LONGLONG' to 'int', possible loss of data 5 d:\tools\Scheme\proj\gambc40b13\lib\os_time.c(235) : warning C4244: 'initializing' : conversion from 'LONGLONG' to 'int', possible loss of data 6 d:\tools\Scheme\proj\gambc40b13\lib\os_time.c(1316) : warning C4113: 'DWORD (__stdcall *)()' differs in parameter lists from 'LPTHREAD_START_ROUTINE' 7 d:\tools\Scheme\proj\gambc40b13\lib\os_shell.c(177) : warning C4244: '=' : conversion from 'unsigned short' to 'char', possible loss of data 8 d:\tools\Scheme\proj\gambc40b13\lib\os_shell.c(444) : warning C4244: '=' : conversion from 'unsigned short' to 'char', possible loss of data 9 d:\tools\Scheme\proj\gambc40b13\lib\os_shell.c(466) : warning C4244: '=' : conversion from 'unsigned short' to 'char', possible loss of data 10 d:\tools\Scheme\proj\gambc40b13\lib\os_shell.c(594) : warning C4244: '=' : conversion from 'unsigned short' to 'char', possible loss of data 11 d:\tools\scheme\proj\gambc40b13\lib\os_io.c(5527) : warning C4700: local variable 'dev' used without having been initialized There are two other warnings that appear about 200 times building the library. I presume they're intentional: 12 d:\tools\Scheme\proj\gambc40b13\lib\_std.c(2011) : warning C4146: unary minus operator applied to unsigned type, result still unsigned 13 d:\tools\Scheme\proj\gambc40b13\lib\_std.c(21221) : warning C4244: '=' : conversion from 'double' to 'float', possible loss of data Here's an example of the first: ___BIGFIX1(-0x80000000L) Here's an example of the second: ___F32VECTORSET(___R1,___R3,___F64V1) ================================================== Problems 1, 2, and 6 are happening because __STDC__ isn't defined by default under MSVC. MS thinks that because they define various extensions they aren't strict ANSI C, so by default they define _MSC_EXTENSIONS instead. Here are the relevant bits from the docs: __STDC__ Indicates full conformance with the ANSI C standard. Defined as the integer constant 1 only if the /Za compiler option is given and you are not compiling C++ code; otherwise is undefined. _MSC_EXTENSIONS This macro is defined when compiling with the /Ze compiler option (the default). Its value, when defined, is 1. Because of all of the above, __P falls through to the #else at 1150 in gambit.h and treats params as K&R, which is wrong. The problem can be fixed by either looking specifically for something like _MFC_VER or _MSC_EXTENSIONS or by forcing a #define of __STDC__. (I took the second route in my project file.) Problem 11 looks legit. (The problem with ___REGISTER_HOST_ENTRY was described elsewhere.)

1 0

Gambit-C 4.0 beta 13
by Marc Feeley 15 Mai '05

15 Mai '05

A new beta of Gambit-C 4.0 is now available in source form at this address: http://www.iro.umontreal.ca/~feeley/gambc40b13.tar.gz Here's what's new: - The "table" type. Tables map keys to values. Gambit has an efficient implementation of tables using hashing and open addressing. The runtime system and compiler now use tables internally instead of association lists in critical places. This improves the speed of the compiler when compiling large files containing many constants, or when the -debug option is used (on some tests the compiler runs 5 times faster). Tables can hold their keys weakly and/or their values weakly. The key comparison procedure and hashing procedure can be specified when a table is created. A reasonably efficient hashing procedure is used by default when the key comparison procedure is eq?, eqv?, equal?, string=? or string-ci=?. - Hashing. There are several hashing procedures (symbol-hash, keyword-hash, string=?-hash, string-ci=?-hash, eq?-hash, eqv?-hash, and equal?-hash). The procedures object->serial-number and serial-number->object are more efficient (thanks to eq? tables). - A new implementation of syntax-case is now included. This version preserves source-code location information, so it makes debugging easier than the previous version. It is usable in the interpreter and in the compiler. The integration with Gambit is not perfect: local variables are renamed, some special forms are transformed (e.g. when pretty-printing a procedure the code may look very different from what the programmer wrote), and some Gambit-specific special forms (such as ##namespace, ##declare, etc) are not available when using syntax-case. For this reason syntax-case is not enabled by default. To use it you must start Gambit like this: % gsi ~~/syntax-case - or load it from your customization file. - The web-server example has been extended to demonstrate how web-continuations can be used. There is also a fairly complete library for dynamically generating HTML. It shows how the ##namespace and ##include forms can be used to modularize code. - Most of the bugs and misfeatures reported on gambit-list have been fixed. Other bugs fixed: equal? on structures, passing Scheme functions to C, imprecise error messages, Mac OS X assembly code problem when using --enable-debug, and more. The system builds cleanly and has been tested in 32 bit and 64 bit environments (mainly Linux and Mac OS X). - The source code for GUIDE (Gambit Universal IDE) is now included in the distribution. Unfortunately, there were some problems getting the makefiles and configure script working properly with Qt, so currently GUIDE is disabled. I expect this to be fixed shortly. [If you feel adventurous the sources are in lib/guide.] - Gambit is now dual licensed. You have the option to choose between the Apache license and the LGPL. Marc

7 21