Re: Trouble compiling with float128 data type

From: Stefano Gragnani (stefano.gragnani_at_[hidden])
Date: 2021-07-21 16:56:00

so how should I complie your example?

and for the following code:

> #include <iomanip>
> #include <iostream>
> #include <boost/multiprecision/cpp_int.hpp>
> #include <boost/multiprecision/cpp_bin_float.hpp>
> #include <boost/multiprecision/float128.hpp>
> #include <boost/math/special_functions/gamma.hpp>
> #include <boost/multiprecision/cpp_dec_float.hpp>
> #include <boost/math/constants/constants.hpp>
> 
> int main()
> {
>    using boost::multiprecision::float128;
> 
>    float128 b = 2;
>    std::cout << std::numeric_limits<float128>::digits << std::endl;
>    std::cout << std::numeric_limits<float128>::digits10 << std::endl;
>    std::cout << std::setprecision(std::numeric_limits<float128>::max_digits10) << log(b) << std::endl;
>    constexpr float128 pi = 3.1415926535897932384626433832795028841971693993751058Q;
>    return 0;
> }   
> 
> I get:
> 
> > Executing task: g++-11 -g -I/usr/local/Cellar/boost/1.76.0/include/ -I/usr/local/include/ -L/usr/local/Cellar/boost/1.76.0/lib/ -L/usr/local/Cellar/gcc/11.1.0_1/lib/gcc/11/ -L/usr/local/Cellar/gmp/6.2.1/lib/ -L/usr/local/Cellar/mpfr/4.1.0/lib/ -std=gnu++20 -lmpfr -lgmp -lquadmath -lboost_system -lboost_filesystem '-fext-numeric-literals' fs_main.cpp -o main <
> 
> Undefined symbols for architecture x86_64:
>   "__Z17quadmath_snprintfPcmPKcz", referenced from:
>       __ZNK5boost14multiprecision8backends16float128_backend3strB5cxx11ElSt13_Ios_Fmtflags in ccGxvsSR.o
>   "__Z4logqg", referenced from:
>       __ZN5boost14multiprecision8backends8eval_logERNS1_16float128_backendERKS2_ in ccGxvsSR.o
> ld: symbol(s) not found for architecture x86_64
> collect2: error: ld returned 1 exit status
> The terminal process "zsh '-c', 'g++-11 -g -I/usr/local/Cellar/boost/1.76.0/include/ -I/usr/local/include/ -L/usr/local/Cellar/boost/1.76.0/lib/ -L/usr/local/Cellar/gcc/11.1.0_1/lib/gcc/11/ -L/usr/local/Cellar/gmp/6.2.1/lib/ -L/usr/local/Cellar/mpfr/4.1.0/lib/ -std=gnu++20 -lmpfr -lgmp -lquadmath -lboost_system -lboost_filesystem '-fext-numeric-literals' fs_main.cpp -o main'" terminated with exit code: 1.

Thanks
Stefano Gragnani

> Il giorno 21 lug 2021, alle ore 17:56, Stefano Gragnani <stefano.gragnani_at_[hidden]> ha scritto:
> 
> ok things are better but i still get errors:
> 
> > Executing task: g++-11 -g -I/usr/local/Cellar/boost/1.76.0/include/ -I/usr/local/include/ -L/usr/local/Cellar/boost/1.76.0/lib/ -L/usr/local/Cellar/gcc/11.1.0_1/lib/gcc/11/ -L/usr/local/Cellar/gmp/6.2.1/lib/ -L/usr/local/Cellar/mpfr/4.1.0/lib/ -std=gnu++20 -lmpfr -lgmp -lquadmath -lboost_system -lboost_filesystem '-fext-numeric-literals' float128_example.cpp -o main <
> 
> float128_example.cpp: In function 'int main()':
> float128_example.cpp:186:20: error: call of overloaded 'exp(__float128)' is ambiguous
>   186 |   float128 e1 = exp(1.Q); // Note argument to exp is type float128.
>       |                 ~~~^~~~~
> In file included from /usr/local/Cellar/gcc/11.1.0_1/include/c++/11.1.0/cmath:45,
>                  from /usr/local/Cellar/boost/1.76.0/include/boost/config/no_tr1/cmath.hpp:21,
>                  from /usr/local/Cellar/boost/1.76.0/include/boost/math/tools/config.hpp:19,
>                  from /usr/local/Cellar/boost/1.76.0/include/boost/math/cstdfloat/cstdfloat_types.hpp:18,
>                  from /usr/local/Cellar/boost/1.76.0/include/boost/cstdfloat.hpp:18,
>                  from float128_example.cpp:19:
> /usr/local/Cellar/gcc/11.1.0_1/lib/gcc/11/gcc/x86_64-apple-darwin20/11.1.0/include-fixed/math.h:376:15: note: candidate: 'double exp(double)'
>   376 | extern double exp(double);
>       |               ^~~
> In file included from /usr/local/Cellar/boost/1.76.0/include/boost/config/no_tr1/cmath.hpp:21,
>                  from /usr/local/Cellar/boost/1.76.0/include/boost/math/tools/config.hpp:19,
>                  from /usr/local/Cellar/boost/1.76.0/include/boost/math/cstdfloat/cstdfloat_types.hpp:18,
>                  from /usr/local/Cellar/boost/1.76.0/include/boost/cstdfloat.hpp:18,
>                  from float128_example.cpp:19:
> /usr/local/Cellar/gcc/11.1.0_1/include/c++/11.1.0/cmath:226:3: note: candidate: 'constexpr long double std::exp(long double)'
>   226 |   exp(long double __x)
>       |   ^~~
> /usr/local/Cellar/gcc/11.1.0_1/include/c++/11.1.0/cmath:222:3: note: candidate: 'constexpr float std::exp(float)'
>   222 |   exp(float __x)
>       |   ^~~
> The terminal process "zsh '-c', 'g++-11 -g -I/usr/local/Cellar/boost/1.76.0/include/ -I/usr/local/include/ -L/usr/local/Cellar/boost/1.76.0/lib/ -L/usr/local/Cellar/gcc/11.1.0_1/lib/gcc/11/ -L/usr/local/Cellar/gmp/6.2.1/lib/ -L/usr/local/Cellar/mpfr/4.1.0/lib/ -std=gnu++20 -lmpfr -lgmp -lquadmath -lboost_system -lboost_filesystem '-fext-numeric-literals' float128_example.cpp -o main'" terminated with exit code: 1.
> 
> Compiling the following example:
> 
> #include <iomanip>
> #include <iostream>
> #include <boost/multiprecision/cpp_int.hpp>
> #include <boost/multiprecision/cpp_bin_float.hpp>
> #include <boost/multiprecision/float128.hpp>
> #include <boost/math/special_functions/gamma.hpp>
> #include <boost/multiprecision/cpp_dec_float.hpp>
> #include <boost/math/constants/constants.hpp>
> 
> int main()
> {
>    using boost::multiprecision::float128;
> 
>    float128 b = 2;
>    std::cout << std::numeric_limits<float128>::digits << std::endl;
>    std::cout << std::numeric_limits<float128>::digits10 << std::endl;
>    std::cout << std::setprecision(std::numeric_limits<float128>::max_digits10) << log(b) << std::endl;
>    constexpr float128 pi = 3.1415926535897932384626433832795028841971693993751058Q;
>    return 0;
> }   
> 
> I get:
> 
> > Executing task: g++-11 -g -I/usr/local/Cellar/boost/1.76.0/include/ -I/usr/local/include/ -L/usr/local/Cellar/boost/1.76.0/lib/ -L/usr/local/Cellar/gcc/11.1.0_1/lib/gcc/11/ -L/usr/local/Cellar/gmp/6.2.1/lib/ -L/usr/local/Cellar/mpfr/4.1.0/lib/ -std=gnu++20 -lmpfr -lgmp -lquadmath -lboost_system -lboost_filesystem '-fext-numeric-literals' fs_main.cpp -o main <
> 
> Undefined symbols for architecture x86_64:
>   "__Z17quadmath_snprintfPcmPKcz", referenced from:
>       __ZNK5boost14multiprecision8backends16float128_backend3strB5cxx11ElSt13_Ios_Fmtflags in ccGxvsSR.o
>   "__Z4logqg", referenced from:
>       __ZN5boost14multiprecision8backends8eval_logERNS1_16float128_backendERKS2_ in ccGxvsSR.o
> ld: symbol(s) not found for architecture x86_64
> collect2: error: ld returned 1 exit status
> The terminal process "zsh '-c', 'g++-11 -g -I/usr/local/Cellar/boost/1.76.0/include/ -I/usr/local/include/ -L/usr/local/Cellar/boost/1.76.0/lib/ -L/usr/local/Cellar/gcc/11.1.0_1/lib/gcc/11/ -L/usr/local/Cellar/gmp/6.2.1/lib/ -L/usr/local/Cellar/mpfr/4.1.0/lib/ -std=gnu++20 -lmpfr -lgmp -lquadmath -lboost_system -lboost_filesystem '-fext-numeric-literals' fs_main.cpp -o main'" terminated with exit code: 1.
> 
> Note that if I compile the following example without using Boost I have no linking errors with quadmath:
> 
> #include <quadmath.h>
> #include <iostream>
> #include <stdlib.h>
> #include <stdio.h>
> 
> int print(const char *label, __float128 r) {
>   int prec = 20;
>   int width = 46;
>   char buf[128];
> 
>   int n = quadmath_snprintf(buf, sizeof buf, "%+-#*.36Qe", width, r);
>   printf ("%s: %s\n", label, buf);
>   return 0;
> }
> 
> int main () {
>   __float128 x = 3.14159265358979323846264338327950288q;
>   print("value", x);
>   print("log", logq(x));
>   print("log10", log10q(x));
>   print("cos", cosq(x));
>   print("sin", sinq(x));
>   print("sqrt", sqrtq(x));
> }
> 
> Executing task: ./es1.out <
> 
> value: +3.141592653589793238462643383279502797e+00   
> log: +1.144729885849400174143427351353058656e+00   
> log10: +4.971498726941338543512682882908988723e-01   
> cos: -1.000000000000000000000000000000000000e+00   
> sin: +8.671810130123781024797044026043352254e-35   
> sqrt: +1.772453850905516027298167483341145140e+00
> 
> The settings are:
> 
> "tasks": [
>       {
>         "label": "Build with g++ 11",
>         "type": "shell",
>         "command": "g++-11",
>         "args": [
>           "-g",
> 	  "-std=gnu++20”,
>           "es1.cpp",
>           "-L/usr/local/Cellar/gcc/11.1.0_1/lib/gcc/11/",
>           "-lquadmath",
>           "-o",
>           "es1.out"
>         ],
>         "group": {
>           "kind": "build",
>           "isDefault": true
>         },
>         "problemMatcher": []
>       },
> 
>> Il giorno 21 lug 2021, alle ore 13:08, John Maddock via Boost-users <boost-users_at_[hidden] <mailto:boost-users_at_[hidden]>> ha scritto:
>> 
>> On 20/07/2021 18:44, Stefano Gragnani via Boost-users wrote:
>>> 
>>> Hi,
>>> 
>>> I'm having trouble compiling code that contains float128 data type.
>>> I'm using MacOS Big Sur 11.4.
>>> Compiler: gcc version 11.1.0 (Homebrew GCC 11.1.0_1)
>>> 
>>> I have no other problems with Boost, the only problem is when I use float128.
>>> 
>>> Please note that I have no problem compiling using libquadmath without using Boost using the __float128 data type.
>>> 
>>> I use Visual Studio Code:
>>> 
>> You can manually fix this by defining BOOST_HAS_FLOAT128 on the command line.  However, you may still hit other errors as __float128 and quadmath.h aren't really fully supported in ISO C++ mode.  The correct fix is to use -std=gnu++20, and then GNU extensions including __float128 are turned on.
>> 
>> HTH, John.
>> 
>> 
>> -- 
>> This email has been checked for viruses by Avast antivirus software.
>> https://www.avast.com/antivirus <https://www.avast.com/antivirus>
>> 
>> _______________________________________________
>> Boost-users mailing list
>> Boost-users_at_[hidden]
>> https://lists.boost.org/mailman/listinfo.cgi/boost-users
>