http://blog.gmane.org/gmane.comp.lib.mpfr.general hourly 1 1901-01-01T00:00+00:00 http://gmane.org/img/gmane-25t.png http://gmane.org http://comments.gmane.org/gmane.comp.lib.mpfr.general/700 <pre>In tsqrt.c, test_property1() claims with a comment (line 497): /* check that -1 &lt;= x/sqrt(x^2+y^2) &lt;= 1 for rounding to nearest or up */ However, the code neglects to use y (except to increment the random number generator): 508 mpfr_urandomb (x, RANDS); 509 mpfr_urandomb (y, RANDS); 510 mpfr_mul (z, x, x, r); 511 mpfr_mul (t, x, x, r); 512 mpfr_add (z, z, t, r); 513 mpfr_sqrt (z, z, r); 514 mpfr_div (z, x, z, r); I think that line 511 is meant to read mpfr_mul(t, y, y, r); Aside: I am only looking at this code in trying to adopt tests for the Ruby bindings for MPFR. </pre> Sam Rawlins 2010-03-10T02:18:18 http://comments.gmane.org/gmane.comp.lib.mpfr.general/696 <pre></pre> Shelley Nation 2010-03-03T04:36:37 http://comments.gmane.org/gmane.comp.lib.mpfr.general/695 <pre>INRIAGforge, which hosts the MPFR repository, will be updated on March 8: https://gforge.inria.fr/forum/forum.php?forum_id=7388 In short, due to this migration, the web pages on INRIAGforge will be unavailable on March 8 from 13:30 to 16:30 CET (12:30 to 15:30 UTC), and access to the Subversion repository will be unavailable from 16:30 to 17:30 CET (15:30 to 16:30 UTC). The MPFR web site www.mpfr.org and mailing-list are not hosted on INRIAGforge, thus will not be affected. Regards, </pre> Vincent Lefevre 2010-03-02T08:32:10 http://comments.gmane.org/gmane.comp.lib.mpfr.general/692 <pre>Hi, I recently installed the mpfr2.4.2 package from the source code.everything while compiling and installing mpfr went well,but after the installation, the libmpfr.so.1 link directing to libmpfr.so.1.2.2 is broken. so every other software which uses libmpfr.so.1 to compile, fails. I also tried recompiling and installing it,but didn't help. Is there any fix for this problem. Thanks in advance. </pre> sirish s 2010-03-01T13:09:00 http://comments.gmane.org/gmane.comp.lib.mpfr.general/691 <pre>As part of the task https://gforge.inria.fr/pm/task.php?func=detailtask&amp;project_task_id=3965&amp;group_id=136&amp;group_project_id=279 I tried to modify the MPFR configure script to reliably recognize long double on PowerPC (with tests on my Mac OS X 10.4.11 machine), which is implemented as a double-double arithmetic. However, due to a bug in GCC (Apple's, but also GCC 4.2.4 and 4.3.4 that are currently installed on my machine), this doesn't work. More precisely, compiling the following program makes cc1 output: tst.c:18: error: initializer for floating value is not a floating constant tst.c:18: confused by earlier errors, bailing out then crash (KERN_PROTECTION_FAILURE). #include &lt;stdio.h&gt; #define E1 (123456789.0L) #define E2 (E1 + 1.0e-60L) long double x = E2; int main (void) { long double y, z; y = E2 - E1; z = x - E1; printf ("%.20g\n", (double) y); printf ("%.20g\n", (double) z); return 0; } I wonder whether it is worth to try to recognize the double-double format as this probably won't be useful in practice (and if I understand correctly, the current solution works with cross-compilation, while an alternative way may not, since it seems that using long double constants doesn't work here). </pre> Vincent Lefevre 2010-02-22T00:48:35 http://comments.gmane.org/gmane.comp.lib.mpfr.general/687 <pre>In reading the MPFR manual, I've noticed that there appears to be no companion function for mpfr_get_z_exp(). mpz_t s; mp_exp_t p; mpfr_t x; p = mpfr_get_z_exp(s, x); for (i = 25; i &lt; 75; i++) mpz_clrbit(s, i); mpfr_set_z_exp(x, s, p, RND); The intent is to manipulate individual bits in the significand. Is there a way to do this that I missed in manual? </pre> Steve Kargl 2010-02-17T23:37:06 http://comments.gmane.org/gmane.comp.lib.mpfr.general/669 <pre>Hello, in the CGAL project, someone recently tried using boost 1.42 and mpfr and noticed compiler errors related to intmax_t. It appears that boost now defines the macros like INTMAX_C when they are not already provided by stdint.h, although it does not add typedefs for intmax_t (not in the global namespace at least). This breaks the intmax_t detection logic of mpfr.h. The platform tested was windows with the microsoft C++ compiler, ie a platform without stdint.h (it looks like version 2010 may finally have it though). Would you consider updating the test in mpfr.h? For instance, you could assume that when _MSC_VER &lt; 1600 there is no intmax_t. I don't know what other platforms may suffer from this. I don't even really know how far mpfr is supporting visual studio (a link to a 3rd party project file?). Last, note that I am posting this following a report by some other user, but I don't have MSVC, so my analysis is questionable. </pre> Marc Glisse 2010-02-14T23:20:03 http://comments.gmane.org/gmane.comp.lib.mpfr.general/664 <pre>I'm working on a Yellowdog Linux Powerstation, trying to upgrade gcc from version 4.1 to the current version. I says I need to upgrade gmp, mpfr, and mpc. The gmp upgrade went fine ( version 4.3.2, not 5.0 ) but, the mpfr configure hangs on this: checking for gfortran... gfortran checking whether we are using the GNU Fortran 77 compiler... this is the configure command output: [root&lt; at &gt;ns1 mpfr-2.4.2]# ./configure --with-gmp-build=/home/danny/Projects/gmp-4.3.2 --prefix=/usr checking for a BSD-compatible install... /usr/bin/install -c checking whether build environment is sane... yes checking for a thread-safe mkdir -p... /bin/mkdir -p checking for gawk... gawk checking whether make sets $(MAKE)... yes checking whether to disable maintainer-specific portions of Makefiles... yes checking build system type... powerpc64-unknown-linux-gnu checking host system type... powerpc64-unknown-linux-gnu checking for grep that handles long lines and -e... /bin/grep checking for egrep... /bin/grep -E checking for a sed that does not truncate output... /bin/sed checking for gcc... gcc checking for C compiler default output file name... a.out checking whether the C compiler works... yes checking whether we are cross compiling... no checking for suffix of executables... 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 ICC... no checking whether gcc and cc understand -c and -o together... yes checking for function prototypes... yes checking for ANSI C header files... yes checking for sys/types.h... yes checking for sys/stat.h... yes checking for stdlib.h... yes checking for string.h... yes checking for memory.h... yes checking for strings.h... yes checking for inttypes.h... yes checking for stdint.h... yes checking for unistd.h... yes checking for string.h... (cached) yes checking for an ANSI C-conforming const... yes checking for working volatile... yes checking for main in -lm... yes checking whether time.h and sys/time.h may both be included... yes checking limits.h usability... yes checking limits.h presence... yes checking for limits.h... yes checking float.h usability... yes checking float.h presence... yes checking for float.h... yes checking for string.h... (cached) yes checking locale.h usability... yes checking locale.h presence... yes checking for locale.h... yes checking wchar.h usability... yes checking wchar.h presence... yes checking for wchar.h... yes checking stdarg.h usability... yes checking stdarg.h presence... yes checking for stdarg.h... yes checking sys/time.h usability... yes checking sys/time.h presence... yes checking for sys/time.h... yes checking sys/fpu.h usability... no checking sys/fpu.h presence... no checking for sys/fpu.h... no checking for stdint.h... (cached) yes checking for SIZE_MAX... yes checking how to copy va_list... va_copy checking for memmove... yes checking for memset... yes checking for setlocale... yes checking for strtol... yes checking for gettimeofday... yes checking for long long int... yes checking for intmax_t... yes checking for union fpc_csr... no checking for fesetround... yes checking for gcc float-conversion bug... no checking for denormalized numbers... yes checking if NAN == NAN... no checking if charset has consecutive values... yes checking for math/round... no checking for math/trunc... no checking for math/floor... yes checking for math/ceil... yes checking for math/rint... no checking for long double... yes checking format of `long double' floating point... unknown (301 235 157 064 124 000 000 000 000 000 000 000 000 000 000 000) checking for gmp.h... yes checking for a sed that does not truncate output... (cached) /bin/sed checking for fgrep... /bin/grep -F checking for ld used by gcc... /usr/bin/ld checking if the linker (/usr/bin/ld) is GNU ld... yes checking for BSD- or MS-compatible name lister (nm)... /usr/bin/nm -B checking the name lister (/usr/bin/nm -B) interface... BSD nm checking whether ln -s works... yes checking the maximum length of command line arguments... 98304 checking whether the shell understands some XSI constructs... yes checking whether the shell understands "+="... yes checking for /usr/bin/ld option to reload object files... -r checking for objdump... objdump checking how to recognize dependent libraries... pass_all checking for ar... ar checking for strip... strip checking for ranlib... ranlib checking command to parse /usr/bin/nm -B output from gcc object... ok checking for dlfcn.h... yes 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 checking for g77... no checking for xlf... no checking for f77... no checking for frt... no checking for pgf77... no checking for cf77... no checking for fort77... no checking for fl32... no checking for af77... no checking for xlf90... no checking for f90... no checking for pgf90... no checking for pghpf... no checking for epcf90... no checking for gfortran... gfortran checking whether we are using the GNU Fortran 77 compiler... I've not been able to find a clue as to why this message. I suspect fortran 77 is not supported but it should just get over it and move on. I don't see anyway to get it to ignore 77 checks other than try to edit the checks out manually in the configure script. I do want fortran support and I would like f77 in the new gcc but can I live without it. Any suggestions would be appreciated. </pre> hembreed 2010-02-04T23:44:21 http://comments.gmane.org/gmane.comp.lib.mpfr.general/659 <pre>Dear Sir, I have a question. Why the addition is not exact? Please see the following codes. x=12388.9, y=12388999, but z=x+y is not 12401387.9. mpfr_t x,y,z; char t[7]={'1','2','3','8','8','.','9'}; char *s=t; char t2[30]={'1','2','3','8','8','9','9','9'}; char *s2=t2; char t1[30]={}; mpfr_init( x); mpfr_set_str(x, s, 10, GMP_RNDN); mpfr_init( y); mpfr_set_str(y, s2, 10, GMP_RNDN); mpfr_init2(z,29); mpfr_add(z,x,y,GMP_RNDN); mp_exp_t exp; mp_exp_t *ex; ex=&amp;exp; mpfr_get_str(t1,ex,10,6,x,GMP_RNDN); cout&lt;&lt;t1&lt;&lt;endl; cout&lt;&lt;*ex&lt;&lt;endl; mpfr_get_str(t1,ex,10,8,y,GMP_RNDN); cout&lt;&lt;t1&lt;&lt;endl; cout&lt;&lt;*ex&lt;&lt;endl; mpfr_get_str(t1,ex,10,10,z,GMP_RNDN); cout&lt;&lt;t1&lt;&lt;endl; cout&lt;&lt;*ex&lt;&lt;endl; mpfr_clear(x); mpfr_clear(y); mpfr_clear(z); mpfr_free_cache(); Sincerely, Shizhong Zhao </pre> zaim-Xb2Y1Dfb9CI< at >public.gmane.org 2010-02-02T19:13:45 http://comments.gmane.org/gmane.comp.lib.mpfr.general/658 <pre>We received the following announce concerning the mpfr.org web site and the MPFR mailing-list: - Common computer equipment and services will be down: from Friday February 5, 5:00 PM to Sunday February 7, 7:00 PM Note: these are CET times, so it would be February 5, 16:00 UTC to February 7, 18:00 UTC. Sorry for the inconvenience, </pre> Vincent Lefevre 2010-01-28T23:55:02 http://comments.gmane.org/gmane.comp.lib.mpfr.general/657 <pre>I've renamed the cumulative patches file "patches" as "allpatches" on the MPFR 2.4.2 page. If you have a link to this file, please update it. This was necessary to avoid a filename clash (yielding an obscure build failure) in case the user downloaded this file into the MPFR source directory under a case-insensitive file system (e.g. under Windows or Mac OS X). Indeed the MPFR source already has a file named "PATCHES". Sorry for the inconvenience, </pre> Vincent Lefevre 2010-01-26T09:50:09 http://comments.gmane.org/gmane.comp.lib.mpfr.general/647 <pre>I just tried running configure on mpfr-2.4.2 using Cygwin. I have installed GMP 5.0.0 using Cygwin. I ran configure with this "./configure --with-gmp-build=/home/chris/gmp-5.0.0". I get this error from configure: "configure: error: BITS_PER_MP_LIMB or BYTES_PER_MP_LIMB are incorrect. You probably need to change some of the GMP or MPFR compile options." I don't understand why I should get this error. Regards Chris Saunders </pre> Chris Saunders 2010-01-24T05:06:03 http://comments.gmane.org/gmane.comp.lib.mpfr.general/641 <pre>I'm trying to build mpfr on a Sun Ultra 27 which is running Open Solaris as the main operating system, but Solaris 10 as a guest operating system inside VirtualBox. bash-3.00$ cat /etc/release Solaris 10 10/09 s10x_u8wos_08a X86 Copyright 2009 Sun Microsystems, Inc. All Rights Reserved. Use is subject to license terms. Assembled 16 September 2009 GMP is installed in /usr/local. I found it necessary to set LDFLAGS to -R /usr/local/lib otherwise the configure script complains that versions of the gmp header and library are different. This is despite the fact I'd used: $ ./configure --with-gmp-lib=/usr/local/lib --with-gmp-include=/usr/local/include Anyway, after I run configure, I can build mpfr with 'make'. However, 'make check' soon fails bash-3.00$ make check Making check in tests make[1]: Entering directory `/export/home/drkirkby/mpfr-2.4.2/tests' make libfrtests.la tversion tinternals tinits tisqrt tsgn tcheck tisnan texceptions tset_exp tset tabs tset_d tset_f tset_q tset_si tset_str tset_z tset_ld tset_sj tswap tcopysign tcmp tcmp2 tcmpabs tcmp_d tcmp_ld tcomparisons teq tadd tsub tmul tdiv tsub1sp tadd1sp tadd_ui tsub_ui tcmp_ui tdiv_ui tmul_ui tsqrt_ui tui_div tui_sub tadd_d tsub_d td_sub tmul_d tdiv_d td_div tgmpop tsi_op tmul_2exp tfma tfms tsum tdim tminmax tnext tfits tget_d tget_d_2exp tget_z tget_str tget_sj tout_str tinp_str toutimpl tcan_round tround_prec tsqrt tconst_log2 tconst_pi tconst_euler trandom ttrunc trint tfrac tmodf texp texp2 texp10 texpm1 tlog tlog2 tlog10 tlog1p tpow tui_pow tpow3 tcosh tsinh ttanh tsinh_cosh tacosh tasinh tatanh thyperbolic tasin tacos tcos tatan tsin ttan tsin_cos tagm thypot tfactorial tgamma terf tcbrt tzeta mpf_compat mpfr_compat reuse tsqr tstrtofr tpow_z tget_f tconst_catalan troot tsec tcsc tcot teint tcoth tcsch tsech tstckintc tsubnormal tlngamma tlgamma tzeta_ui tget_ld_2exp tget_set_d64 tj0 tj1 tjn ty0 ty1 tyn tremquo tfmod tl2b tli2 tprintf tsprintf tfprintf trec_sqrt tpow_all make[2]: Entering directory `/export/home/drkirkby/mpfr-2.4.2/tests' make[2]: `libfrtests.la' is up to date. /bin/bash ../libtool --tag=CC --mode=link gcc -Wall -Wmissing-prototypes -Wpointer-arith -m32 -O2 -fomit-frame-pointer -mtune=nocona -march=nocona -ffloat-store -R /usr/local/lib -L/usr/local/lib -o tversion tversion.o libfrtests.la -lm ../libmpfr.la -lgmp libtool: link: gcc -Wall -Wmissing-prototypes -Wpointer-arith -m32 -O2 -fomit-frame-pointer -mtune=nocona -march=nocona -ffloat-store -o .libs/tversion tversion.o -L/usr/local/lib ./.libs/libfrtests.a -lm ../.libs/libmpfr.so /usr/local/lib/libgmp.so -R/usr/local/lib Undefined first referenced symbol in file mpz_div_ui ../.libs/libmpfr.so mpz_div_2exp ../.libs/libmpfr.so ld: fatal: Symbol referencing errors. No output written to .libs/tversion collect2: ld returned 1 exit status make[2]: *** [tversion] Error 1 make[2]: Leaving directory `/export/home/drkirkby/mpfr-2.4.2/tests' make[1]: *** [check-am] Error 2 make[1]: Leaving directory `/export/home/drkirkby/mpfr-2.4.2/tests' make: *** [check-recursive] Error 1 Any ideas? I've attached config.log Dave </pre> Dr. David Kirkby 2010-01-20T13:09:39 http://comments.gmane.org/gmane.comp.lib.mpfr.general/640 <pre>In the repository I was appearing under two usernames: lefevre (which came from the old CVS repository) and vlefevre. I've changed all the "lefevre" into "vlefevre". If you use a mirror (e.g. with svnsync), these revision properties won't necessarily be updated (for instance, you probably need svnsync copy-revprops ... to update them). </pre> Vincent Lefevre 2010-01-19T02:17:41 http://comments.gmane.org/gmane.comp.lib.mpfr.general/620 <pre> I'm trying to build mpfr 2.4.2 but I get the error in 'make check' that is described in http://www.mpfr.org/faq.html#undef_ref1 I've configured my env as recommended to point to /usr/local/include and /usr/local/lib # export C_INCLUDE_PATH=/usr/local/include # export LD_LIBRARY_PATH=/usr/local/lib # export LIBRARY_PATH=/usr/local/lib # export LD_RUN_PATH=/usr/local/lib # ls $C_INCLUDE_PATH gmp.h # ls $LD_LIBRARY_PATH libgmp.a libgmp.la* libgmp.so&lt; at &gt; libgmp.so.3&lt; at &gt; libgmp.so.3.5.2* # ls $LIBRARY_PATH libgmp.a libgmp.la* libgmp.so&lt; at &gt; libgmp.so.3&lt; at &gt; libgmp.so.3.5.2* this looks OK. 'make clean &amp;&amp; make' works (I get a clean build) but 'make check' still reports a problem: Making check in tests make[1]: Entering directory `/r/sanyo.unx.sas.com/vol/vol220/u22/sasdjb/dev/openmp/gomp/mpfr-2.4.2/tests' make libfrtests.la tversion tinternals tinits tisqrt tsgn tcheck tisnan texceptions tset_exp tset tabs tset_d tset_f tset_q tset_si tset_str tset_z tset_ld tset_sj tswap tcopysign tcmp tcmp2 tcmpabs tcmp_d tcmp_ld tcomparisons teq tadd tsub tmul tdiv tsub1sp tadd1sp tadd_ui tsub_ui tcmp_ui tdiv_ui tmul_ui tsqrt_ui tui_div tui_sub tadd_d tsub_d td_sub tmul_d tdiv_d td_div tgmpop tsi_op tmul_2exp tfma tfms tsum tdim tminmax tnext tfits tget_d tget_d_2exp tget_z tget_str tget_sj tout_str tinp_str toutimpl tcan_round tround_prec tsqrt tconst_log2 tconst_pi tconst_euler trandom ttrunc trint tfrac tmodf texp texp2 texp10 texpm1 tlog tlog2 tlog10 tlog1p tpow tui_pow tpow3 tcosh tsinh ttanh tsinh_cosh tacosh tasinh tatanh thyperbolic tasin tacos tcos tatan tsin ttan tsin_cos tagm thypot tfactor ial tgamma terf tcbrt tzeta mpf_compat mpfr_compat reuse tsqr tstrtofr tpow_z tget_f tconst_catalan troot tsec tcsc tcot teint tcoth tcsch tsech tstckintc tsubnormal tlngamma tlgamma tzeta_u i tget_ld_2exp tget_set_d64 tj0 tj1 tjn ty0 ty1 tyn tremquo tfmod tl2b tli2 tprintf tsprintf tfprintf trec_sqrt tpow_all make[2]: Entering directory `/r/sanyo.unx.sas.com/vol/vol220/u22/sasdjb/dev/openmp/gomp/mpfr-2.4.2/tests' make[2]: `libfrtests.la' is up to date. /bin/sh ../libtool --tag=CC --mode=link gcc -std=gnu99 -Wall -Wmissing-prototypes -Wpointer-arith -O2 -pedantic -m64 -mtune=k8 -o tversion tversion.o libfrtests.la -lm ../libmpfr.la -lgmp libtool: link: gcc -std=gnu99 -Wall -Wmissing-prototypes -Wpointer-arith -O2 -pedantic -m64 -mtune=k8 -o .libs/tversion tversion.o ./.libs/libfrtests.a -lm ../.libs/libmpfr.so -lgmp -Wl,-rpath -Wl,/usr/local/lib ./.libs/libfrtests.a(tests.o): In function `src_fopen': tests.c:(.text+0x51e): undefined reference to `__gmp_get_memory_functions' tests.c:(.text+0x581): undefined reference to `__gmp_get_memory_functions' collect2: ld returned 1 exit status make[2]: *** [tversion] Error 1 make[2]: Leaving directory `/r/sanyo.unx.sas.com/vol/vol220/u22/sasdjb/dev/openmp/gomp/mpfr-2.4.2/tests' make[1]: *** [check-am] Error 2 make[1]: Leaving directory `/r/sanyo.unx.sas.com/vol/vol220/u22/sasdjb/dev/openmp/gomp/mpfr-2.4.2/tests' make: *** [check-recursive] Error 1 If instead I do make clean ./configure --with-gmp=/usr/local make make check then the tests all work correctly. However, as stated in the FAQ, this is less desirable since I have to configure other software to do the same. Should the build work with these environment variable settings? if so, how can I make it work? My environment: # uname -a Linux acl811.unx.sas.com 2.6.18-128.el5 #1 SMP Wed Dec 17 11:41:38 EST 2008 x86_64 x86_64 x86_64 GNU/Linux thanks! </pre> David J. Biesack 2010-01-12T20:04:08 http://comments.gmane.org/gmane.comp.lib.mpfr.general/615 <pre> Hi, with GMP 5.0.0 and MPFR, it is possible to compute 10^9 bits of Pi in less than one hour on a 2.83Ghz Core 2: patate% time ./tconst_pi 1000000000 0 0 Seed GMP_CHECK_RANDOMIZE=1263667646 (include this in bug reports) 2885.349u 192.229s 52:16.55 98.1% 0+0k 280+16io 3pf+0w Paul Zimmermann </pre> Paul Zimmermann 2010-01-12T14:34:21 http://comments.gmane.org/gmane.comp.lib.mpfr.general/614 <pre>What would it take to implement random distributions according to the philosophy of MPFR? Consider sampling from a Gaussian with mean mu and variance sigma^2: It would be nice if this were equivalent to sampling from the exact (continuous) distribution and rounding the result (according to the rounding direction) to the closest representable number. We would stipulate, of course, a source of random bits (which in practise would come from some high-quality random number generator). I've attempted to follow this philosophy in RandomLib http://randomlib.sourceforge.net as applied to the standard floating data types. In particular, this offers "exact" implementations of the uniform distribution in [0,1] and [-1,1] and the outcome of Bernoulli trials. (Here "exact" assumes a source of random bits.) Some observations: * In addition to rounding to the nearest representable number, it's probably useful to offer a mode where the result is rounded to a multiple of a fixed interval. (This may make the analysis of subsequent operations easier.) * Even if p/q is not representable, it's possible to implement Bernoulli trials with probability p/q exactly. http://randomlib.sourceforge.net/html/RandomCanonical_8hpp-source.html#l00939 * It takes, on average, only 2 bits to implement Bernoulli trials with probability p (a float, double, or long double). http://randomlib.sourceforge.net/html/RandomCanonical_8hpp-source.html#l00884 * It's possible to implement Bernoulli trials with probability exp(-p) exactly using a technique due to von Neumann. http://randomlib.sourceforge.net/html/classRandomLib_1_1ExponentialProb.html * Similarly it's possible to sample from exp(-x) exactly. Essentially random numbers allow you to compute log(U) exactly using just comparisons. http://randomlib.sourceforge.net/html/classRandomLib_1_1ExactExponential.html * It's not obvious (to me, at any rate) how to generalize the method used to sample from exp(-x) to sampling from a Gaussian (or most other non-trivial distributions for that matter). * It's difficult to conceive of simulations where this sort of accuracy would be noticeable. Despite this, there's surely a theoretical interest in this sort of numerical problem. --Charles </pre> Karney, Charles 2010-01-12T02:29:46 http://comments.gmane.org/gmane.comp.lib.mpfr.general/609 <pre> Environment: AlphaServer ES40 model2 # psrinfo -v Status of processor 0 as of: 01/08/10 08:59:10 Processor has been on-line since 01/01/2010 14:53:51 The alpha EV6.8AL (21264B) processor operates at 834 MHz, has a cache size of 8388608 bytes, and has an alpha internal floating point processor. Status of processor 1 as of: 01/08/10 08:59:10 Processor has been on-line since 01/01/2010 14:53:51 The alpha EV6.8AL (21264B) processor operates at 834 MHz, has a cache size of 8388608 bytes, and has an alpha internal floating point processor. Status of processor 2 as of: 01/08/10 08:59:10 Processor has been on-line since 01/01/2010 14:53:51 The alpha EV6.8AL (21264B) processor operates at 834 MHz, has a cache size of 8388608 bytes, and has an alpha internal floating point processor. Status of processor 3 as of: 01/08/10 08:59:10 Processor has been on-line since 01/01/2010 14:53:51 The alpha EV6.8AL (21264B) processor operates at 834 MHz, has a cache size of 8388608 bytes, and has an alpha internal floating point processor. # vmstat -P Total Physical Memory = 16384.00 M = 2097152 pages # OS # ./config.guess alphaev68-dec-osf5.1b - mpre precisely Tru64unix version 5.1B-5 (BL28) # cc -V Compaq C V6.5-011 on HP Tru64 UNIX V5.1B (Rev. 2650) Compiler Driver V6.5-003 (sys) cc Driver Trying to build mpfr-2.4.2 over recently released gmp distributions... mpfr-2.4.2 configures and builds without problems using gmp-4.3.2. After building gmp-5.0.0, which built and checked without errors, building mpfr-2.4.2 using --with-gmp-build option produces the following mpfr-2.4.2 ./configure output: # ./configure --with-gmp-build=/usr/local/JunqueYard/gmp-5.0.0 checking for a BSD-compatible install... ./install-sh -c checking whether build environment is sane... yes checking for a thread-safe mkdir -p... ./install-sh -c -d checking for gawk... gawk checking whether make sets $(MAKE)... yes checking whether to disable maintainer-specific portions of Makefiles... yes checking build system type... alphaev68-dec-osf5.1b checking host system type... alphaev68-dec-osf5.1b checking for grep that handles long lines and -e... /usr/local/bin/grep checking for egrep... /usr/local/bin/grep -E checking for a sed that does not truncate output... /usr/local/bin/sed checking for CC=cc and CFLAGS= -fast -arch ev67 -tune ev67... yes checking for gcc... cc checking for C compiler default output file name... a.out checking whether the C compiler works... yes checking whether we are cross compiling... no checking for suffix of executables... checking for suffix of object files... o checking whether we are using the GNU C compiler... no checking whether cc accepts -g... yes checking for cc option to accept ISO C89... none needed checking for style of include used by make... GNU checking dependency style of cc... tru64 checking how to run the C preprocessor... cc -E checking for ICC... no checking whether cc understands -c and -o together... yes checking for function prototypes... yes checking for ANSI C header files... yes checking for sys/types.h... yes checking for sys/stat.h... yes checking for stdlib.h... yes checking for string.h... yes checking for memory.h... yes checking for strings.h... yes checking for inttypes.h... yes checking for stdint.h... no checking for unistd.h... yes checking for string.h... (cached) yes checking for an ANSI C-conforming const... yes checking for working volatile... yes checking for main in -lm... yes checking whether time.h and sys/time.h may both be included... yes checking limits.h usability... yes checking limits.h presence... yes checking for limits.h... yes checking float.h usability... yes checking float.h presence... yes checking for float.h... yes checking for string.h... (cached) yes checking locale.h usability... yes checking locale.h presence... yes checking for locale.h... yes checking wchar.h usability... yes checking wchar.h presence... yes checking for wchar.h... yes checking stdarg.h usability... yes checking stdarg.h presence... yes checking for stdarg.h... yes checking sys/time.h usability... yes checking sys/time.h presence... yes checking for sys/time.h... yes checking sys/fpu.h usability... no checking sys/fpu.h presence... no checking for sys/fpu.h... no checking for stdint.h... (cached) no checking for SIZE_MAX... (((1UL &lt;&lt; 63) - 1) * 2 + 1) checking how to copy va_list... memcpy checking for memmove... yes checking for memset... yes checking for setlocale... yes checking for strtol... yes checking for gettimeofday... yes checking for IEEE-754 switches... -fprm d -ieee_with_inexact checking for long long int... yes checking for intmax_t... no checking for union fpc_csr... no checking for fesetround... no checking for denormalized numbers... yes checking if NAN == NAN... no checking if charset has consecutive values... yes checking for math/round... no checking for math/trunc... yes checking for math/floor... yes checking for math/ceil... yes checking for math/rint... no checking for long double... yes checking format of `long double' floating point... unknown (000 000 000 000 000 000 000 000 000 000 100 105 363 326 031 300) checking for gmp.h... yes checking for a sed that does not truncate output... (cached) /usr/local/bin/sed checking for fgrep... /usr/local/bin/grep -F checking for non-GNU ld... /usr/bin/ld checking if the linker (/usr/bin/ld) is GNU ld... no checking for BSD- or MS-compatible name lister (nm)... /usr/bin/nm -B checking the name lister (/usr/bin/nm -B) interface... BSD nm checking whether ln -s works... yes checking the maximum length of command line arguments... 16384 checking whether the shell understands some XSI constructs... yes checking whether the shell understands "+="... no checking for /usr/bin/ld option to reload object files... -r checking for objdump... objdump checking how to recognize dependent libraries... pass_all checking for ar... ar checking for strip... strip checking for ranlib... ranlib checking command to parse /usr/bin/nm -B output from cc object... ok checking for dlfcn.h... yes checking for g++... no checking for c++... no checking for gpp... no checking for aCC... no checking for CC... no checking for cxx... cxx checking whether we are using the GNU C++ compiler... no checking whether cxx accepts -g... yes checking dependency style of cxx... tru64 checking how to run the C++ preprocessor... cxx -E checking for g77... no checking for xlf... no checking for f77... no checking for frt... no checking for pgf77... no checking for cf77... no checking for fort77... no checking for fl32... no checking for af77... no checking for xlf90... no checking for f90... no checking for pgf90... no checking for pghpf... no checking for epcf90... no checking for gfortran... no checking for g95... no checking for xlf95... no checking for f95... no checking for fort... no checking for ifort... no checking for ifc... no checking for efc... no checking for pgf95... no checking for lf95... no checking for ftn... no checking whether we are using the GNU Fortran 77 compiler... no checking whether accepts -g... no checking for objdir... .libs checking for cc option to produce PIC... -DPIC checking if cc PIC flag -DPIC works... yes checking if cc static flag -non_shared works... yes checking if cc supports -c -o file.o... yes checking if cc supports -c -o file.o... (cached) yes checking whether the cc linker (/usr/bin/ld) supports shared libraries... yes checking dynamic linker characteristics... osf5.1b ld.so checking how to hardcode library paths into programs... immediate checking whether stripping libraries is possible... no checking if libtool supports shared libraries... yes checking whether to build shared libraries... yes checking whether to build static libraries... yes checking whether the cxx linker (/usr/bin/ld) supports shared libraries... yes checking for cxx option to produce PIC... -DPIC checking if cxx PIC flag -DPIC works... yes checking if cxx static flag -non_shared works... yes checking if cxx supports -c -o file.o... yes checking if cxx supports -c -o file.o... (cached) yes checking whether the cxx linker (/usr/bin/ld) supports shared libraries... yes checking dynamic linker characteristics... osf5.1b ld.so checking how to hardcode library paths into programs... immediate checking whether gcc __attribute__ ((mode (XX))) works... no checking for recent GMP... yes checking for gmp internal files... yes checking for valid BITS_PER_MP_LIMB and BYTES_PER_MP_LIMB... no configure: error: BITS_PER_MP_LIMB or BYTES_PER_MP_LIMB are incorrect. You probably need to change some of the GMP or MPFR compile options. # Can you suggest which GMP or MPFR compile options may be necessary? Thank you, Bill Glessner Central Washington University </pre> Bill.Glessner-3+CQIGFm+IQ< at >public.gmane.org 2010-01-08T17:22:06 http://comments.gmane.org/gmane.comp.lib.mpfr.general/604 <pre>----- Forwarded message from noreply-QteCRy7xEqLvuXihggDW+w&lt; at &gt;public.gmane.org ----- From: noreply-QteCRy7xEqLvuXihggDW+w&lt; at &gt;public.gmane.org Subject: [INRIAGforge] :: Erratum :: January 2010, the 14th : scheduled service downtime Date: Thu, 07 Jan 2010 19:59:35 +0100 Due to sheduled maintenance, the INRIA Forge will be shut down from 17h00. The restart is planned from 19h00 on the same day. We apologize for any inconvenience. Cordially, -- The InriaGforge Team ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ 14 Janvier 2010 (et non pas 14 Octobre 2009) : arrêt du service pour maintenance La Forge INRIA sera arrêtée à partir de 17h00 pour de maintenance. Le redémarrage est prévu à partir de 19h00 le jour même. Nous sommes désolés pour le dérangement occasionné. Cordialement, -- L'équipe InriaGforge ----- End forwarded message ----- </pre> Vincent Lefevre 2010-01-07T23:45:32 http://comments.gmane.org/gmane.comp.lib.mpfr.general/598 <pre>Hello! Here is a testsuite failure of successfully built MPFR 2.4.2 with latest (-p2) patches applied, and linked against GMP 4.3.1 built just before it, under the same environment. Environment specs follow, testsuite failure log is at the bottom: # sizer -v HP Tru64 UNIX V5.1B (Rev. 2650); Thu Dec 17 18:47:21 CET 2009 # psrinfo -v 0 Status of processor 0 as of: 12/28/09 23:43:00 Processor has been on-line since 12/28/2009 17:30:17 The alpha EV5.6 (21164A) processor operates at 468 MHz, has a cache size of 4194304 bytes, and has an alpha internal floating point processor. # psrinfo -n number of processors on system = 4 # cc -V Compaq C V6.5-303 (dtk) on HP Tru64 UNIX V5.1B (Rev. 2650) Compiler Driver V6.5-302 (dtk) cc Driver # cxx -V Compaq C++ V7.1-006 for HP Tru64 UNIX V5.1B (Rev. 2650) Compiler Driver V7.1-006 (cxx) cxx Driver # pwd /usr/src/gcc/mpfr-2.4.2 # ./config.guess alphaev56-dec-osf5.1b # ./configure --prefix=/opt/gcc --with-gmp=/opt/gcc [...] # gmake -j6 [...] # gmake -j6 check [...] gmake check-TESTS gmake[2]: Entering directory `/usr/src/gcc/mpfr-2.4.2/tests' PASS: tversion PASS: tinternals PASS: tinits PASS: tisqrt PASS: tsgn PASS: tcheck PASS: tisnan PASS: texceptions PASS: tset_exp PASS: tset PASS: tabs PASS: tset_d PASS: tset_f PASS: tset_q PASS: tset_si PASS: tset_str PASS: tset_z PASS: tset_ld PASS: tset_sj PASS: tswap PASS: tcopysign PASS: tcmp PASS: tcmp2 PASS: tcmpabs PASS: tcmp_d PASS: tcmp_ld PASS: tcomparisons PASS: teq PASS: tadd PASS: tsub PASS: tmul PASS: tdiv PASS: tsub1sp PASS: tadd1sp PASS: tadd_ui PASS: tsub_ui PASS: tcmp_ui PASS: tdiv_ui PASS: tmul_ui PASS: tsqrt_ui PASS: tui_div PASS: tui_sub PASS: tadd_d PASS: tsub_d PASS: td_sub PASS: tmul_d PASS: tdiv_d PASS: td_div PASS: tgmpop PASS: tsi_op PASS: tmul_2exp PASS: tfma PASS: tfms PASS: tsum PASS: tdim PASS: tminmax PASS: tnext PASS: tfits PASS: tget_d PASS: tget_d_2exp PASS: tget_z PASS: tget_str PASS: tget_sj PASS: tout_str PASS: tinp_str PASS: toutimpl PASS: tcan_round PASS: tround_prec PASS: tsqrt PASS: tconst_log2 PASS: tconst_pi PASS: tconst_euler PASS: trandom PASS: ttrunc PASS: trint PASS: tfrac PASS: tmodf PASS: texp PASS: texp2 PASS: texp10 PASS: texpm1 PASS: tlog PASS: tlog2 PASS: tlog10 PASS: tlog1p PASS: tpow PASS: tui_pow PASS: tpow3 PASS: tcosh PASS: tsinh PASS: ttanh PASS: tsinh_cosh PASS: tacosh PASS: tasinh PASS: tatanh PASS: thyperbolic PASS: tasin PASS: tacos PASS: tcos PASS: tatan PASS: tsin PASS: ttan PASS: tsin_cos PASS: tagm PASS: thypot PASS: tfactorial PASS: tgamma PASS: terf PASS: tcbrt PASS: tzeta PASS: mpf_compat PASS: mpfr_compat PASS: reuse PASS: tsqr PASS: tstrtofr PASS: tpow_z PASS: tget_f PASS: tconst_catalan PASS: troot PASS: tsec PASS: tcsc PASS: tcot PASS: teint PASS: tcoth PASS: tcsch PASS: tsech PASS: tstckintc PASS: tsubnormal PASS: tlngamma PASS: tlgamma PASS: tzeta_ui PASS: tget_ld_2exp PASS: tget_set_d64 PASS: tj0 PASS: tj1 PASS: tjn PASS: ty0 PASS: ty1 PASS: tyn PASS: tremquo PASS: tfmod PASS: tl2b PASS: tli2 Error in test #8: mpfr_printf printed 21 characters instead of 20 FAIL: tprintf PASS: tsprintf Error in test #8: mpfr_vfprintf printed 21 characters instead of 20 FAIL: tfprintf PASS: trec_sqrt PASS: tpow_all ===================== 2 of 148 tests failed ===================== gmake[2]: *** [check-TESTS] Error 1 gmake[2]: Leaving directory `/usr/src/gcc/mpfr-2.4.2/tests' gmake[1]: *** [check-am] Error 2 gmake[1]: Leaving directory `/usr/src/gcc/mpfr-2.4.2/tests' gmake: *** [check-recursive] Error 1 # Please note that I'm not subscribed to the list, so please CC me explicitly with any replies. Thanks! Regards, Sab </pre> RUMI Szabolcs 2009-12-29T02:34:16 http://comments.gmane.org/gmane.comp.lib.mpfr.general/592 <pre>Is exp(x) broken in mpfr or am I misunstanding how mpfr works? The function (exp(x) - 1) / x goes to one as x goes to zero. In the code below, I start at x = 1 and successively divide by 2. #include &lt;stdio.h&gt; #include "mpfr.h" #include "gmp.h" int main(void) { mpfr_t x, e; int i; mpfr_init2(x,53); mpfr_init2(e,53); mpfr_set_ui(x, 1, GMP_RNDN); for (i = 0; i &lt; 80; i++) { mpfr_exp(e, x, GMP_RNDN); mpfr_sub_ui(e, e, 1, GMP_RNDN); mpfr_div(e, e, x, GMP_RNDN); mpfr_printf("%d %.15Re %.15Re\n", i, x, e); mpfr_div_ui(x, x, 2, GMP_RNDN); } } Here is select output: 23 1.192092895507812e-07 1.000000059604645e+00 24 5.960464477539062e-08 1.000000029802322e+00 25 2.980232238769531e-08 1.000000014901161e+00 &lt;-- 26 1.490116119384766e-08 1.000000014901161e+00 &lt;-- This looks suspicious 27 7.450580596923828e-09 1.000000000000000e+00 &lt;-- 28 3.725290298461914e-09 1.000000000000000e+00 ... 51 4.440892098500626e-16 1.000000000000000e+00 52 2.220446049250313e-16 1.000000000000000e+00 53 1.110223024625157e-16 2.000000000000000e+00 &lt;-- This looks suspicious 54 5.551115123125783e-17 0.000000000000000e+00 &lt;-- 55 2.775557561562891e-17 0.000000000000000e+00 Am I'm doing something dumb? </pre> Steve Kargl 2009-12-19T23:40:04 Search the mailing list at Gmane query http://search.gmane.org/?group=$group=gmane.comp.lib.mpfr.general