- CLN-list - ginac.de

Re: [CLN-list] CLN
by Richard B. Kreckel 08 Nov '10

08 Nov '10

Dear Kåre, On 10/29/2010 09:33 PM, Kåre Olaussen wrote: > Dear Bruno and Richard. > > I tried to subscribe to the CLN mailing list, but the subscription page does not work. Indeed. There is a problem in openSUSE 11.3 packaging leading to spurious error messages after subscribing. (The version of MailMan in openSUSE 11.3 is less than the one in previous versions of openSUSE and not up-to-date for Python 2.6 shipped with openSUSE 11.*.) This will be fixed some time. Anyway, you are subscribed now, according to the members list. I'm CC'ing you just to be sure. > We want to use CLN to make world record calculations in various directions, > but this requires effective use of parallel (or multi-thread) computations. > > Our experience with (careful) multithread computations using CLN is spotless. > I have read your reservation in the mailing list archive that CLN is not 'thread-safe'. > > I don't understand what you mean by that! All CLN objects are reference-counted and this code is susceptible to race-conditions where different threads manipulate the reference counts without paying attention to concurrent threads. > Is that like declaring that "no car is collision safe" (in which sense I think no software > package is 'thread-safe'), or are there specific problem areas you know of? > We have already made many-many-million CLN algebraic operations both in > serial and threaded (using openMP), and always obtained the same results. That is entirely possible if you don't share CLN objects between threads or do so only very infrequently. The most likely problem you'll ever hit is probably a segmentation fault while accessing memory another thread as prematurely freed. > We also want to run our (CLN) programs on large computer clusters -- using MPI. > This could be made much more efficient if we could transfer for instance CL_F > numbers between different computers in binary form. > > Could you please help me understand how this can be done? Instead of hacking something low-level, let me make a suggestion: convert your floating-point numbers to a triple (sign, mantissa, exponent), serialize it to a buffer in base 32, send that across your cluster and let the receiver reverse this operation. Check the manual for the type decoded_float. The point is that conversion to/from a base that is a power of two is O(N) with little overhead. It is significantly more efficient than conversion to any other base which is O(N*log(N)*log(log(N))). And 32 is the largest supported base. This way, the overhead compared to the internal representation is only 1-log2(32)/log2(256)=3/8. Unless your program is extremely communication-intense (in which case you should think about redesigning it), this is almost optimal. Best wishes -richy. -- Richard B. Kreckel <http://www.ginac.de/~kreckel/>

1 0

Linker error: multiple definition of...
by Lorne McIntosh 06 Oct '10

06 Oct '10

Hi all, I'm attempting to build CLN 1.3.1 on Windows 7 32-bit using MinGW (and NetBeans IDE 6.9.1). The build seemed to be going well until linking, when I got the following error: ../src/.libs/libcln.a(cl_F_decode.o):cl_F_decode.cc:(.text$_ZN3cln12decode_floatERKNS_5cl_LFE[cln::decode_float(cln::cl_LF const&)]+0x0): multiple definition of `cln::decode_float(cln::cl_LF const&)' ../src/.libs/libcln.a(cl_LF_decode.o):d:\School\Grad\Thesis\Prototype09\cln-1.3.1\src/float/lfloat/misc/cl_LF_decode.cc:19: first defined here

2 1

Difficulty building rpm of CLN
by galen 22 Sep '10

22 Sep '10

I am working on an updated rpm of your library for PCLinuxOS. I was hoping for some assistance, as I am running out of ideas. I appreciated any attention you may offer to this problem. Thanks for this library. First of all, I can build the library successfully using ./configure, make, and make install. However, when I try to build it with a spec file, it comes up with all kinds of errors. I started with the spec PCLinuxOS spec file for 1.1.13, then looked for clues from Mandriva, Fedora and the included cln.spec.in. Here seems to be the environment created for building the rpm: + CFLAGS='-O2 -pipe -Wp,-D_FORTIFY_SOURCE=2 -fomit-frame-pointer -march=i586 -mtune=generic -fasynchronous-unwind-tables' + export CFLAGS + CXXFLAGS='-O2 -pipe -Wp,-D_FORTIFY_SOURCE=2 -fomit-frame-pointer -march=i586 -mtune=generic -fasynchronous-unwind-tables' + export CXXFLAGS + FFLAGS='-O2 -pipe -Wp,-D_FORTIFY_SOURCE=2 -fomit-frame-pointer -march=i586 -mtune=generic -fasynchronous-unwind-tables' Here are the errors when I build without any other modifications: checking whether build environment is sane... /usr/share/automake-1.11/am/depend2.am: am__fastdepCCAS does not appear in AM_CONDITIONAL /usr/share/automake-1.11/am/depend2.am: The usual way to define `am__fastdepCCAS' is to add `AM_PROG_AS' /usr/share/automake-1.11/am/depend2.am: to `configure.ac' and run `aclocal' and `autoconf' again. src/Makefile.am: Preprocessed Assembler source seen but `CCAS' is undefined src/Makefile.am: The usual way to define `CCAS' is to add `AM_PROG_AS' src/Makefile.am: to `configure.ac' and run `aclocal' and `autoconf' again. src/Makefile.am: Preprocessed Assembler source seen but `CCASFLAGS' is undefined src/Makefile.am: The usual way to define `CCASFLAGS' is to add `AM_PROG_AS' src/Makefile.am: to `configure.ac' and run `aclocal' and `autoconf' again. yes checking for a thread-safe mkdir -p... /bin/mkdir -p checking for gawk... gawk checking whether make sets $(MAKE)... yes checking for gcc... gcc checking whether the C compiler works... yes checking for C compiler default output file name... a.out checking for suffix of executables... checking whether we are cross compiling... no checking for suffix of object files... o checking whether we are using the GNU C compiler... yes checking whether gcc accepts -g... yes checking for gcc option to accept ISO C89... none needed checking for style of include used by make... GNU checking dependency style of gcc... gcc3 checking how to run the C preprocessor... gcc -E checking for g++... g++ checking whether we are using the GNU C++ compiler... yes checking whether g++ accepts -g... yes checking dependency style of g++... gcc3 checking how to run the C++ preprocessor... g++ -E ./configure: line 4759: CL_AS_UNDERSCORE: command not found ./configure: line 4760: CL_AS_NOEXECSTACK: command not found ./configure: line 4762: syntax error near unexpected token `autoconf' ./configure: line 4762: ` CL_CANONICAL_HOST(autoconf)' make: *** [config.status] Error 2 make: *** Waiting for unfinished jobs.... make: *** [Makefile.in] Error 1 error: Bad exit status from /home/galen/src/tmp/rpm-tmp.68019 (%build) RPM build errors: Bad exit status from /home/galen/src/tmp/rpm-tmp.68019 (%build) From this output, I've tried adding: echo 'AM_PROG_AS' >> configure.ac aclocal autoconf just before %configure. This leads to other errors which I'll gladly outline if necessary. Thank you for taking the time to read this, and any feedback would be appreciated. Galen

2 1

Conversion bug?
by Jens Vollinga 21 Sep '10

21 Sep '10

Hi, is there maybe a bug in the implementation of double_approx()? If I do something like #include <iostream> #include <cln/cln.h> using namespace std; using namespace cln; int main() { cl_F c = cl_F("1.23456789012345678901234567890_200"); cout << c << endl; cout << double_approx(c) << endl; return 0; } the second output gives only single precision (aka float_approx)! Regards, Jens

2 1

CLN patch 2
by Jan private 11 Sep '10

11 Sep '10

Hello Alexei, here is another hairy one. 1. Problem: MSVC does not tag "extern cl_private_thing cl_I_constructor_from_UL (uint32 wert);" in cln/number.h with the cln namespace because it is declared inside the body of an inlined function (symbol created is >>?cl_I_constructor_from_UL@@YAPAUcl_anything@cln@@K@Z<<) But cl_I_constructor_from_UL is defined inside cln namespace in cl_I_from_UL.cc (symbol created is >>?cl_I_constructor_from_UL@cln@@YAPAUcl_anything@1@K@Z<<) So we get unresolved external symbols on every assignment and constructor operation. 2. Solution (rather, work-around): Create an auxiliary function cln::cl_private_thing cl_I_constructor_from_UL (uint32 wert); outside of namespace cln which references cln::cl_private_thing cln::cl_I_constructor_from_UL(wert); Note: The attached patch is exemplary only. As soon as the method to solve this issue is agreed on, similar patches will have to be made for other constructors Best regards, Jan

1 0

swapping numbers
by Dejan Jovanović 26 Aug '10

26 Aug '10

Hi, I'm writing an application with intense usage of integers and one of the operations that might have a performance impact is swapping two numbers. I don't know the internals of the CLN library and I see there is no swap operation in the headers. Could someone please elaborate how much stuff happens internally when I do the usual swap: void swap(cl_I& a, cl_I& b) { (1) cl_I temp; (2) temp = a; (3) a = b; (4) b = temp; (5) } My worst-case guess is: (1) a sequence of constructors (2) reconstruction of a (3) destruction of a, reconstruction of b (4) destruction of b, reconstruction of temp (5) destruction of temp Also, would massive amount of swaps lead to internal memory fragmentation due to construction/destruction? Thanks, Dejan

3 5

how to read string numbers at runtime into cl_I type
by Joshua Friedman 21 Jul '10

21 Jul '10

I have a text file like this: 16 8534483450280192 17 236164889075956560 18 -2445879764812789440 19 14227278615873217344 ... Integer paired with a very large integer. How do I read in from a file to store the big integer in a cl_I type. I tried reading it in as a string, and then I couldn't convert from a string to any of the types (cl_R, or cl_I). Eventually I'll need it to be a float, which I should be able to use cl_float() to convert from an exact type. Thanks Josh -- Joshua Friedman PhD CrownEagle(a)gmail.com http://www.math.sunysb.edu/~joshua

3 3

Compile with MSVC successful
by Jan Rheinlaender 21 Jul '10

21 Jul '10

I successfully compiled CLN-1.3.1 with Microsoft Visual C++. You can find an account of my adventures on http://sourceforge.net/apps/mediawiki/ooo-imath/index.php?title=Installatio… The benchmarks, examples and tests compile and both tests claim to pass successfully. Since I am not an expert programmer, I have no idea if my source code modifications are acceptable or "evil". Have fun, Jan

1 0

file input and cln number types
by Joshua Friedman 15 Jul '10

15 Jul '10

I have a text file like this: 16 8534483450280192 17 236164889075956560 18 -2445879764812789440 19 14227278615873217344 ... Integer paired with a very large integer. How do I read in from a file to store the big integer in a cl_I type. I tried reading it in as a string, and then I couldn't convert from a string to any of the types (cl_R, or cl_I). Eventually I'll need it to be a float, which I should be able to use cl_float() to convert from an exact type. Thanks Josh -- Joshua Friedman PhD CrownEagle(a)gmail.com http://www.math.sunysb.edu/~joshua

1 0

[PATCH 2/2] Configure: be more cross-compilation friendly.
by Alexei Sheplyakov 28 May '10

28 May '10

Rewrite CL_GMP_SET_UINTD so it works for cross-compilation too. --- m4/gmp.m4 | 67 ++++++++++++++++++++++++++++++++++++------------------------ 1 files changed, 40 insertions(+), 27 deletions(-) diff --git a/m4/gmp.m4 b/m4/gmp.m4 index e0e5ef3..455c200 100644 --- a/m4/gmp.m4 +++ b/m4/gmp.m4 @@ -36,33 +36,46 @@ cl_cv_new_libgmp="yes", cl_cv_new_libgmp="no") dnl What is sizeof(mp_limb_t)? (It has to match sizeof(uintD) later.) AC_DEFUN([CL_GMP_SET_UINTD], [AC_CACHE_CHECK([how large gmp demands uintD to be], cl_cv_gmp_set_uintd, [ - AC_TRY_RUN([#include <gmp.h> -#include <stdio.h> -int main() { - FILE *f=fopen("conftestval", "w"); - if (!f) return(255); - if (sizeof(mp_limb_t) > sizeof(long)) - fprintf(f, "long long"); - else if (sizeof(mp_limb_t) == sizeof(long)) - fprintf(f, "long"); - else if (sizeof(mp_limb_t) == sizeof(int)) - fprintf(f, "int"); - else return(sizeof(mp_limb_t)); -#if defined(__GMP_BITS_PER_MP_LIMB) - /* Is there a nail in a limb? */ - if (8*sizeof(mp_limb_t)!=__GMP_BITS_PER_MP_LIMB) - return(254); -#endif - return(0); -}], cl_cv_gmp_set_uintd=`cat conftestval` - cl_gmp_demands="GMP_DEMANDS_UINTD_`echo ${cl_cv_gmp_set_uintd} | sed -e 'y/ gilnot/_GILNOT/'`", - gmp_retval="$ac_status" - if test x$gmp_retval = "x255"; then AC_MSG_ERROR([error opening output file.]); fi - if test x$gmp_retval = "x254"; then AC_MSG_ERROR([nails in MP limbs are unsupported.]); fi - AC_MSG_ERROR([Don't know which C-type has sizeof $gmp_retval.]), - AC_MSG_ERROR([cross-compiling - cannot determine])) -]) -AC_DEFINE_UNQUOTED($cl_gmp_demands) + dnl Note: we don't run any of compiled programs here, so this method + dnl both works for native and cross compilation + cl_gmp_demands="UNKNOWN" + cl_gmp_has_nails="no" + AC_COMPILE_IFELSE([AC_LANG_PROGRAM([[#include <gmp.h> + template<bool COND> struct Static_Assert; + template<> struct Static_Assert<true> { }; + #if defined(__GMP_BITS_PER_MP_LIMB) + Static_Assert<8*sizeof(mp_limb_t) == __GMP_BITS_PER_MP_LIMB> check; + #endif]], [[]])], [], [cl_gmp_has_nails="yes"]) + if test "x$cl_gmp_has_nails" = "xyes"; then + AC_MSG_ERROR([nails in MP libms are unsupported.]) + fi + AC_COMPILE_IFELSE([AC_LANG_PROGRAM([[#include <gmp.h> + template<bool COND> struct Static_Assert; + template<> struct Static_Assert<true> { }; + Static_Assert<sizeof(mp_limb_t) > sizeof(long)> check;]], [[]])], + [cl_gmp_demands='GMP_DEMANDS_UINTD_LONG_LONG'], []) + if test "x$cl_gmp_demands" = "xUNKNOWN"; then + AC_COMPILE_IFELSE([AC_LANG_PROGRAM([[#include <gmp.h> + template<bool COND> struct Static_Assert; + template<> struct Static_Assert<true> { }; + Static_Assert<sizeof(mp_limb_t) == sizeof(long)> check;]], [[]])], + [cl_gmp_demands='GMP_DEMANDS_UINTD_LONG'], []) + fi + if test "x$cl_gmp_demands" = "xUNKNOWN"; then + AC_COMPILE_IFELSE([AC_LANG_PROGRAM([[ + #include <gmp.h> + template<bool COND> struct Static_Assert; + template<> struct Static_Assert<true> { }; + Static_Assert<sizeof(mp_limb_t) == sizeof(int)> check;]], [[]])], + [cl_gmp_demands='GMP_DEMANDS_UINTD_INT'], []) + fi + if test "x$cl_gmp_demands" = "xUNKNOWN"; then + AC_MSG_ERROR([Don't know which C-type has sizeof(mp_limb_t)]) + else + cl_cv_gmp_set_uintd="$cl_gmp_demands" + fi + ]) +AC_DEFINE_UNQUOTED($cl_cv_gmp_set_uintd) ]) dnl Whether or not to use GMP. Sets CL_USE_GMP. -- 1.7.1

1 0