int128 on Linux for Intel compiler

alexvi · ‎03-13-2013

Linux 2.6.32

Intel compiler: icpc version 13.0.1 (gcc version 4.4.6 compatibility)

#include <iostream> #include <sys/types.h> int main() { std::cerr << sizeof (__uint128_t) << std::endl; return 0; } Output: 16

So, type __uint128_t exists. However, working with __uint128_t produces compilation and run errors.

Programs:

------------------ File int01.cpp #include <cstdint> int main() { uint128_t val128 = 0; return 0; }

------------------ File int02.cpp #include <cstdint> int main() { __uint128_t val128 = 0; return 0; }

------------------ File int03.cpp #include <iostream> #include <cstdint> int main() { __uint128_t val128 = 0; std::cerr << val128 << std::endl; return 0; }

Compilations:

> icpc int01.cpp /usr/include/c++/4.4.6/c++0x_warning.h(31): catastrophic error: #error directive: This file requires compiler and library support for the upcoming ISO C++ standard, C++0x. This support is currently experimental, and must be enabled with the -std=c++0x or -std=gnu++0x compiler options. #error This file requires compiler and library support for the upcoming \ ^

compilation aborted for int01.cpp (code 4)

> icpc int01.cpp -std=c++0x int01.cpp(4): error: identifier "uint128_t" is undefined uint128_t val128 = 0; ^

compilation aborted for int01.cpp (code 2)

> icpc int02.cpp -std=c++0x // No errors

> icpc int03.cpp -std=c++0x

int03.cpp(6): error: more than one operator "<<" matches these operands: function "std::basic_ostream<_CharT, _Traits>::operator<<(long) [with _CharT=char, _Traits=std::char_traits<char>]" function "std::basic_ostream<_CharT, _Traits>::operator<<(unsigned long) [with _CharT=char, _Traits=std::char_traits<char>]" function "std::basic_ostream<_CharT, _Traits>::operator<<(bool) [with _CharT=char, _Traits=std::char_traits<char>]" function "std::basic_ostream<_CharT, _Traits>::operator<<(short) [with _CharT=char, _Traits=std::char_traits<char>]" function "std::basic_ostream<_CharT, _Traits>::operator<<(unsigned short) [with _CharT=char, _Traits=std::char_traits<char>]" function "std::basic_ostream<_CharT, _Traits>::operator<<(int) [with _CharT=char, _Traits=std::char_traits<char>]" function "std::basic_ostream<_CharT, _Traits>::operator<<(unsigned int) [with _CharT=char, _Traits=std::char_traits<char>]" function "std::basic_ostream<_CharT, _Traits>::operator<<(long long) [with _CharT=char, _Traits=std::char_traits<char>]" function "std::basic_ostream<_CharT, _Traits>::operator<<(unsigned long long) [with _CharT=char, _Traits=std::char_traits<char>]" function "std::basic_ostream<_CharT, _Traits>::operator<<(double) [with _CharT=char, _Traits=std::char_traits<char>]" function "std::basic_ostream<_CharT, _Traits>::operator<<(float) [with _CharT=char, _Traits=std::char_traits<char>]" function "std::basic_ostream<_CharT, _Traits>::operator<<(long double) [with _CharT=char, _Traits=std::char_traits<char>]" operand types are: std::ostream << __uint128_t std::cerr << val128 << std::endl; ^

compilation aborted for int03.cpp (code 2)

SergeyKostrov · ‎03-13-2013

>>... icpc int01.cpp /usr/include/c++/4.4.6/c++0x_warning.h(31): catastrophic error: #error directive: This file requires... Could you attach c++0x_warning.h(31) and cstdint header files? >>...int01.cpp(4): error: identifier "uint128_t" is undefined uint128_t val128 = 0;... There are no underscores for uint128_t in that case.

SergeyKostrov · ‎03-13-2013

>>...File int03.cpp #include #include int main() { __uint128_t val128 = 0; std::cerr << val128... I see that there are No C++ operator << for the type __uint128_t and this is why the last test failed and try to use printf CRT function instead. The following types could be used with C++ operator <<: ::operator << (long) ::operator << (unsigned long) ::operator << (bool) ::operator << (short) ::operator << (unsigned short) ::operator << (int) ::operator << (unsigned int) ::operator << (long long) ::operator << (unsigned long long) ::operator << (double) ::operator << (float) ::operator << (long double)

alexvi · ‎03-13-2013

CRT function: what does it mean?

alexvi · ‎03-13-2013

CRT function: what does it mean?

alexvi · ‎03-13-2013

Attached

alexvi · ‎03-13-2013

Attached

alexvi · ‎03-13-2013

Attached

SergeyKostrov · ‎03-13-2013

>>...CRT function: what does it mean? C Run-Time. All functions, like printf, fprintf, etc, are called CRT functions.

SergeyKostrov · ‎03-13-2013

Define __GXX_EXPERIMENTAL_CXX0X__ macro at the beginning of your sources.

alexvi · ‎03-13-2013

printf (..., "%llx", ) doesn't work for uint128.

Could you send a small program with uint128 that prints uint128-value?

Thanks

SergeyKostrov · ‎03-13-2013

>>...Could you attach ... cstdint header files? I'd like to see how the type __uint128_t is declared in cstdint header file.

alexvi · ‎03-13-2013

__GXX_EXPERIMENTAL_CXX0X__ #include <iostream> #include <cstdint> int main() { __uint128_t val128 = 0; std::cerr << val128 << std::endl; return 0; }

> icpc int03.cpp

Log attached

alexvi · ‎03-13-2013

// <cstdint> -*- C++ -*-

// Copyright (C) 2007, 2008, 2009 Free Software Foundation, Inc. // // This file is part of the GNU ISO C++ Library. This library is free // software; you can redistribute it and/or modify it under the // terms of the GNU General Public License as published by the // Free Software Foundation; either version 3, or (at your option) // any later version.

// This library is distributed in the hope that it will be useful, // but WITHOUT ANY WARRANTY; without even the implied warranty of // MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the // GNU General Public License for more details.

// Under Section 7 of GPL version 3, you are granted additional // permissions described in the GCC Runtime Library Exception, version // 3.1, as published by the Free Software Foundation.

// You should have received a copy of the GNU General Public License and // a copy of the GCC Runtime Library Exception along with this program; // see the files COPYING3 and COPYING.RUNTIME respectively. If not, see // <http://www.gnu.org/licenses/>.

/** @file include/cstdint * This is a Standard C++ Library header. */

#ifndef _GLIBCXX_CSTDINT #define _GLIBCXX_CSTDINT 1

#pragma GCC system_header

#ifndef __GXX_EXPERIMENTAL_CXX0X__ # include <c++0x_warning.h> #endif

#if defined(_GLIBCXX_INCLUDE_AS_TR1) # error C++0x header cannot be included from TR1 header #endif

#include <bits/c++config.h>

// For 8.22.1/1 (see C99, Notes 219, 220, 222) #if _GLIBCXX_HAVE_STDINT_H # ifndef __STDC_LIMIT_MACROS # define _UNDEF__STDC_LIMIT_MACROS # define __STDC_LIMIT_MACROS # endif # ifndef __STDC_CONSTANT_MACROS # define _UNDEF__STDC_CONSTANT_MACROS # define __STDC_CONSTANT_MACROS # endif # include <stdint.h> # ifdef _UNDEF__STDC_LIMIT_MACROS # undef __STDC_LIMIT_MACROS # undef _UNDEF__STDC_LIMIT_MACROS # endif # ifdef _UNDEF__STDC_CONSTANT_MACROS # undef __STDC_CONSTANT_MACROS # undef _UNDEF__STDC_CONSTANT_MACROS # endif #endif

#if defined(_GLIBCXX_INCLUDE_AS_CXX0X) # include <tr1_impl/cstdint> #else # define _GLIBCXX_INCLUDE_AS_CXX0X # define _GLIBCXX_BEGIN_NAMESPACE_TR1 # define _GLIBCXX_END_NAMESPACE_TR1 # define _GLIBCXX_TR1 # include <tr1_impl/cstdint> # undef _GLIBCXX_TR1 # undef _GLIBCXX_END_NAMESPACE_TR1 # undef _GLIBCXX_BEGIN_NAMESPACE_TR1 # undef _GLIBCXX_INCLUDE_AS_CXX0X #endif

#endif // _GLIBCXX_CSTDINT

SergeyKostrov · ‎03-13-2013

>>...File int02.cpp #include int main() { __uint128_t val128 = 0; return 0; } #include int main() { __uint128_t val128 = 12345678901234567890ULL; return 0; } For the second test use modulus val128 % 10 to print all the digits in the 128-bit number. You need to implement a simple for loop with 128 iterations with printf which prints every digit. Unfortunately, I don't see a builtin support for __uint128_t type in printf CRT-function.

alexvi · ‎03-13-2013

Thanks.

But problem is not only with print.

For instance, std::numeric_limits<__uint128_>::max() returns 0.

SergeyKostrov · ‎03-13-2013

>>...You need to implement a simple for loop with 128 iterations with printf which prints every digit.... Correction: The for loop should work until ( val128 % 10 ) is not equal to zero. >>...For instance, std::numeric_limits<__uint128_>::max() returns 0... It means, that support for that type is partially implemented.

alexvi · ‎03-13-2013

#include <iostream> #include <iomanip> #include <cstdint> int main() { __uint128_t val128 = 0xeeee1234567812345678; unsigned long val64l = static_cast<unsigned long>((val128 << 64) >> 64); unsigned long val64m = static_cast<unsigned long>(val128 >> 64); std::cerr << std::hex << val64l << std::dec << std::endl; std::cerr << std::hex << val64m << std::dec << std::endl; return 0; }

> icpc int03.cpp -std=c++0x

> ./a.out

560800000000 // Expected 1234567812345678

eeee

SergeyKostrov · ‎03-13-2013

>>>>...For instance, std::numeric_limits<__uint128_>::max() returns 0... >> >>It means, that support for that type is partially implemented. Do you need this in a real application or just for tests? Take a look at how std::numeric_limits<__uint128_>::max() is implemented and add support. It will be your own patch / fix ( 3.4028236692093846346337460743177e+38 - 1 ).

alexvi · ‎03-13-2013

I need it in a real application.

But "partially supported" feature is worse than non-supported one.

SergeyKostrov · ‎03-13-2013

>>For instance, std::numeric_limits[ __uint128_ ]::max() returns 0... Did you miss t in __uint128_? >>...# include "stdint.h"... Could you attach stdint.h header file? Sorry, but I still don't see how __uint128_t is declared. In Intel C++ compiler v12.x there is no support for that type ( just verified ). I'll take a look at v13.x and let you know.