[Boost-bugs] [Boost C++ Libraries] #9717: Boost math library on PPC64 has thousands of errors

Subject: [Boost-bugs] [Boost C++ Libraries] #9717: Boost math library on PPC64 has thousands of errors
From: Boost C++ Libraries (noreply_at_[hidden])
Date: 2014-02-26 14:36:01


#9717: Boost math library on PPC64 has thousands of errors
-------------------------------------+-------------------------
 Reporter: Tony Reix <tony.reix@…> | Owner: johnmaddock
     Type: Bugs | Status: new
Milestone: To Be Determined | Component: math
  Version: Boost 1.55.0 | Severity: Problem
 Keywords: |
-------------------------------------+-------------------------
 Hi,

 I'm trying to port Boost 1.55 on PowerPC64 BE and LE.
 Thousands of errors appear with math library.

 Investigation one test (bessel_zeros_example_1) which loops, I see by
 means of gdb the following:
   boost::math::cyl_bessel_j_zero< double, .....
   boost::math::detail::cyl_bessel_j_zero_imp<__float128, .....

 This "__float128" seems erroneous.
 However, looking at boost/math/special_functions/bessel.hpp code, I see
 that this type is chosen at compilation time, by means of:
  typedef typename policies::evaluation<result_type, Policy>::type
 value_type;
 and it is a nightmare to understand why the compiler did choose
 __float128. Unless gdb is showing nuts ?

 Exact test is:
 bin.v2/libs/math/example/bessel_zeros_example_1.test/gcc-4.8.2/debug/bessel_zeros_example_1

 Comparing execution path of this test code in PPC64/BE to Intel64, I see
 some floor() value computed differently on the 2 systems, and leading the
 test program to enter a wrong path and stay stucked in an infinite loop.

 There is some fundamental error, hidden by compiler which compiles
 templates and gives me no hint.

 I need help to determine what is wrong with Boost math on PPC64.

 Tony

-- 
Ticket URL: <https://svn.boost.org/trac/boost/ticket/9717>
Boost C++ Libraries <http://www.boost.org/>
Boost provides free peer-reviewed portable C++ source libraries.

This archive was generated by hypermail 2.1.7 : 2017-02-16 18:50:15 UTC