September 2009 - PPL-devel - BUGSENG Mailman

[GIT] ppl/ppl(floating_point): Added Fabio Bossi to the list of current contributors.
by Roberto Bagnara 02 Sep '09

02 Sep '09

Module: ppl/ppl Branch: floating_point Commit: 091ba8a954912a4eb80447877629c9882f449901 URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=091ba8a95491… Author: Roberto Bagnara <bagnara(a)cs.unipr.it> Date: Wed Sep 2 15:46:12 2009 +0200 Added Fabio Bossi to the list of current contributors. --- CREDITS | 4 ++++ 1 files changed, 4 insertions(+), 0 deletions(-) diff --git a/CREDITS b/CREDITS index 7316d9b..2f76e24 100644 --- a/CREDITS +++ b/CREDITS @@ -35,6 +35,10 @@ Current Contributors: intervals. He also helps on several other design and implementation issues. + Fabio Bossi (student of the University of Parma) + is working on the PPL support for the approximation + of floating point computations. + Andrea Cimino (former student of the University of Parma) wrote most of the mixed integer programming solver, and also most of the Java and OCaml

1 0

[GIT] ppl/ppl(floating_point): Removed dummy linearization operators.
by Fabio Bossi 02 Sep '09

02 Sep '09

Module: ppl/ppl Branch: floating_point Commit: 12d9fe8700455e6b66f5389e5bf3778573debfb4 URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=12d9fe870045… Author: Fabio Bossi <bossi(a)cs.unipr.it> Date: Wed Sep 2 16:33:29 2009 +0200 Removed dummy linearization operators. Implemented the linearization operator for Opposite_Floating_Point_Expression. --- src/Division_Floating_Point_Expression.inlines.hh | 10 ---------- src/Makefile.am | 5 ++++- ...iplication_Floating_Point_Expression.inlines.hh | 12 ------------ src/Opposite_Floating_Point_Expression.defs.hh | 2 +- src/Opposite_Floating_Point_Expression.inlines.hh | 2 +- 5 files changed, 6 insertions(+), 25 deletions(-) diff --git a/src/Division_Floating_Point_Expression.inlines.hh b/src/Division_Floating_Point_Expression.inlines.hh index 0b92afb..0d0c96b 100644 --- a/src/Division_Floating_Point_Expression.inlines.hh +++ b/src/Division_Floating_Point_Expression.inlines.hh @@ -54,16 +54,6 @@ Division_Floating_Point_Expression<FP_Interval_Type, FP_Format> std::swap(second_operand, y.second_operand); } - // FIXME: add function body -template <typename FP_Interval_Type, typename FP_Format> -inline -typename Division_Floating_Point_Expression<FP_Interval_Type, FP_Format> -::FP_Linear_Form Division_Floating_Point_Expression<FP_Interval_Type, FP_Format> -::linearize(const FP_Interval_Abstract_Store& store) const { - FP_Linear_Form result = FP_Linear_Form(); - return result; -} - } // namespace Parma_Polyhedra_Library namespace std { diff --git a/src/Makefile.am b/src/Makefile.am index 90a3c86..a12639d 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -358,7 +358,10 @@ Sum_Floating_Point_Expression.templates.hh \ Difference_Floating_Point_Expression.types.hh \ Difference_Floating_Point_Expression.defs.hh \ Difference_Floating_Point_Expression.inlines.hh \ -Difference_Floating_Point_Expression.templates.hh +Difference_Floating_Point_Expression.templates.hh \ +Opposite_Floating_Point_Expression.types.hh \ +Opposite_Floating_Point_Expression.defs.hh \ +Opposite_Floating_Point_Expression.inlines.hh #Any_Pointset.types.hh \ #Any_Pointset.defs.hh \ #Any_Pointset.inlines.hh \ diff --git a/src/Multiplication_Floating_Point_Expression.inlines.hh b/src/Multiplication_Floating_Point_Expression.inlines.hh index 1784476..f0206bd 100644 --- a/src/Multiplication_Floating_Point_Expression.inlines.hh +++ b/src/Multiplication_Floating_Point_Expression.inlines.hh @@ -56,18 +56,6 @@ Multiplication_Floating_Point_Expression<FP_Interval_Type, FP_Format> std::swap(second_operand, y.second_operand); } - - // FIXME: add function body -template <typename FP_Interval_Type, typename FP_Format> -inline -typename Multiplication_Floating_Point_Expression<FP_Interval_Type,FP_Format> -::FP_Linear_Form -Multiplication_Floating_Point_Expression<FP_Interval_Type, FP_Format> -::linearize(const FP_Interval_Abstract_Store& store) const { - FP_Linear_Form result = FP_Linear_Form(); - return result; -} - } // namespace Parma_Polyhedra_Library namespace std { diff --git a/src/Opposite_Floating_Point_Expression.defs.hh b/src/Opposite_Floating_Point_Expression.defs.hh index 536f4c3..2a04cac 100644 --- a/src/Opposite_Floating_Point_Expression.defs.hh +++ b/src/Opposite_Floating_Point_Expression.defs.hh @@ -43,7 +43,7 @@ namespace Parma_Polyhedra_Library { template <typename FP_Interval_Type, typename FP_Format> class Opposite_Floating_Point_Expression -: public Floating_Point_Expression<FP_Interval_Type, FP_Format> { + : public Floating_Point_Expression<FP_Interval_Type, FP_Format> { public: diff --git a/src/Opposite_Floating_Point_Expression.inlines.hh b/src/Opposite_Floating_Point_Expression.inlines.hh index 0e59f37..a99ea05 100644 --- a/src/Opposite_Floating_Point_Expression.inlines.hh +++ b/src/Opposite_Floating_Point_Expression.inlines.hh @@ -56,7 +56,7 @@ inline typename Opposite_Floating_Point_Expression<FP_Interval_Type, FP_Format>::FP_Linear_Form Opposite_Floating_Point_Expression<FP_Interval_Type, FP_Format> ::linearize(const FP_Interval_Abstract_Store& store) const { - // FIXME: implement this. + return -(operand->linearize(store)). } } // namespace Parma_Polyhedra_Library

1 0

[GIT] ppl/ppl(floating_point): Added a few more implementations of the linearization operator.
by Fabio Bossi 02 Sep '09

02 Sep '09

Module: ppl/ppl Branch: floating_point Commit: fd45e56cd50f609ddb815735f487576349ac6a07 URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=fd45e56cd50f… Author: Fabio Bossi <bossi(a)cs.unipr.it> Date: Wed Sep 2 16:02:45 2009 +0200 Added a few more implementations of the linearization operator. --- src/Difference_Floating_Point_Expression.defs.hh | 1 + ...Difference_Floating_Point_Expression.inlines.hh | 11 ---- ...fference_Floating_Point_Expression.templates.hh | 51 ++++++++++++++++++++ src/Makefile.am | 10 ++++- src/Sum_Floating_Point_Expression.defs.hh | 1 + src/Sum_Floating_Point_Expression.inlines.hh | 11 ---- src/Sum_Floating_Point_Expression.templates.hh | 50 +++++++++++++++++++ 7 files changed, 112 insertions(+), 23 deletions(-) diff --git a/src/Difference_Floating_Point_Expression.defs.hh b/src/Difference_Floating_Point_Expression.defs.hh index f2c0ad6..304e92e 100644 --- a/src/Difference_Floating_Point_Expression.defs.hh +++ b/src/Difference_Floating_Point_Expression.defs.hh @@ -100,5 +100,6 @@ private: } // namespace Parma_Polyhedra_Library #include "Difference_Floating_Point_Expression.inlines.hh" +#include "Difference_Floating_Point_Expression.templates.hh" #endif // !defined(PPL_Difference_Floating_Point_Expression_defs_hh) diff --git a/src/Difference_Floating_Point_Expression.inlines.hh b/src/Difference_Floating_Point_Expression.inlines.hh index 744207d..596b21e 100644 --- a/src/Difference_Floating_Point_Expression.inlines.hh +++ b/src/Difference_Floating_Point_Expression.inlines.hh @@ -55,17 +55,6 @@ Difference_Floating_Point_Expression<FP_Interval_Type, FP_Format> std::swap(second_operand, y.second_operand); } - // FIXME: add function body -template <typename FP_Interval_Type, typename FP_Format> -inline -typename Difference_Floating_Point_Expression<FP_Interval_Type, FP_Format> -::FP_Linear_Form -Difference_Floating_Point_Expression<FP_Interval_Type, FP_Format> -::linearize(const FP_Interval_Abstract_Store& store) const { - FP_Linear_Form result = FP_Linear_Form(); - return result; -} - } // namespace Parma_Polyhedra_Library namespace std { diff --git a/src/Difference_Floating_Point_Expression.templates.hh b/src/Difference_Floating_Point_Expression.templates.hh new file mode 100644 index 0000000..24eee4e --- /dev/null +++ b/src/Difference_Floating_Point_Expression.templates.hh @@ -0,0 +1,51 @@ +/* Difference_Floating_Point_Expression class implementation: + non-inline template functions. + Copyright (C) 2001-2009 Roberto Bagnara <bagnara(a)cs.unipr.it> + +This file is part of the Parma Polyhedra Library (PPL). + +The PPL 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 of the License, or (at your +option) any later version. + +The PPL 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. + +You should have received a copy of the GNU General Public License +along with this program; if not, write to the Free Software Foundation, +Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02111-1307, USA. + +For the most up-to-date information see the Parma Polyhedra Library +site: http://www.cs.unipr.it/ppl/ . */ + +#ifndef PPL_Difference_Floating_Point_Expression_templates_hh +#define PPL_Difference_Floating_Point_Expression_templates_hh 1 + +#include "Difference_Floating_Point_Expression.defs.hh" + +namespace Parma_Polyhedra_Library { + +template <typename FP_Interval_Type, typename FP_Format> +typename Difference_Floating_Point_Expression<FP_Interval_Type, FP_Format> +::FP_Linear_Form Difference_Floating_Point_Expression<FP_Interval_Type, + FP_Format> +::linearize(const FP_Interval_Abstract_Store& store) const { + FP_Linear_Form linearized_first_operand = first_operand->linearize(store); + FP_Linear_Form linearized_second_operand = second_operand->linearize(store); + FP_Interval error = FP_Interval_Type(-absolute_error); + // FIXME: this may be incorrect for some policies. + error.join_assign(absolute_error); + FP_Linear_Form abs_error = FP_Linear_Form(error); + FP_Linear_Form result = linearized_first_operand - linearized_second_operand + + relative_error(linearized_first_operand) + + relative_error(linearized_second_operand) + + abs_error; + return result; +} + +} // namespace Parma_Polyhedra_Library + +#endif // !defined(PPL_Difference_Floating_Point_Expression_templates_hh) diff --git a/src/Makefile.am b/src/Makefile.am index 9fd2582..90a3c86 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -350,7 +350,15 @@ Constant_Floating_Point_Expression.defs.hh \ Constant_Floating_Point_Expression.inlines.hh \ Variable_Floating_Point_Expression.types.hh \ Variable_Floating_Point_Expression.defs.hh \ -Variable_Floating_Point_Expression.inlines.hh +Variable_Floating_Point_Expression.inlines.hh \ +Sum_Floating_Point_Expression.types.hh \ +Sum_Floating_Point_Expression.defs.hh \ +Sum_Floating_Point_Expression.inlines.hh \ +Sum_Floating_Point_Expression.templates.hh \ +Difference_Floating_Point_Expression.types.hh \ +Difference_Floating_Point_Expression.defs.hh \ +Difference_Floating_Point_Expression.inlines.hh \ +Difference_Floating_Point_Expression.templates.hh #Any_Pointset.types.hh \ #Any_Pointset.defs.hh \ #Any_Pointset.inlines.hh \ diff --git a/src/Sum_Floating_Point_Expression.defs.hh b/src/Sum_Floating_Point_Expression.defs.hh index eec9115..57671ef 100644 --- a/src/Sum_Floating_Point_Expression.defs.hh +++ b/src/Sum_Floating_Point_Expression.defs.hh @@ -97,5 +97,6 @@ private: } // namespace Parma_Polyhedra_Library #include "Sum_Floating_Point_Expression.inlines.hh" +#include "Sum_Floating_Point_Expression.templates.hh" #endif // !defined(PPL_Sum_Floating_Point_Expression_defs_hh) diff --git a/src/Sum_Floating_Point_Expression.inlines.hh b/src/Sum_Floating_Point_Expression.inlines.hh index 7051f16..aa73ba9 100644 --- a/src/Sum_Floating_Point_Expression.inlines.hh +++ b/src/Sum_Floating_Point_Expression.inlines.hh @@ -55,17 +55,6 @@ Sum_Floating_Point_Expression<FP_Interval_Type, FP_Format> std::swap(second_operand, y.second_operand); } - - // FIXME: add function body -template <typename FP_Interval_Type, typename FP_Format> -inline -typename Sum_Floating_Point_Expression<FP_Interval_Type, FP_Format> -::FP_Linear_Form Sum_Floating_Point_Expression<FP_Interval_Type, FP_Format> -::linearize(const FP_Interval_Abstract_Store& store) const { - FP_Linear_Form result = FP_Linear_Form(); - return result; -} - } // namespace Parma_Polyhedra_Library namespace std { diff --git a/src/Sum_Floating_Point_Expression.templates.hh b/src/Sum_Floating_Point_Expression.templates.hh new file mode 100644 index 0000000..0ac587a --- /dev/null +++ b/src/Sum_Floating_Point_Expression.templates.hh @@ -0,0 +1,50 @@ +/* Sum_Floating_Point_Expression class implementation: + non-inline template functions. + Copyright (C) 2001-2009 Roberto Bagnara <bagnara(a)cs.unipr.it> + +This file is part of the Parma Polyhedra Library (PPL). + +The PPL 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 of the License, or (at your +option) any later version. + +The PPL 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. + +You should have received a copy of the GNU General Public License +along with this program; if not, write to the Free Software Foundation, +Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02111-1307, USA. + +For the most up-to-date information see the Parma Polyhedra Library +site: http://www.cs.unipr.it/ppl/ . */ + +#ifndef PPL_Sum_Floating_Point_Expression_templates_hh +#define PPL_Sum_Floating_Point_Expression_templates_hh 1 + +#include "Sum_Floating_Point_Expression.defs.hh" + +namespace Parma_Polyhedra_Library { + +template <typename FP_Interval_Type, typename FP_Format> +typename Sum_Floating_Point_Expression<FP_Interval_Type, FP_Format> +::FP_Linear_Form Sum_Floating_Point_Expression<FP_Interval_Type, FP_Format> +::linearize(const FP_Interval_Abstract_Store& store) const { + FP_Linear_Form linearized_first_operand = first_operand->linearize(store); + FP_Linear_Form linearized_second_operand = second_operand->linearize(store); + FP_Interval error = FP_Interval_Type(-absolute_error); + // FIXME: this may be incorrect for some policies. + error.join_assign(absolute_error); + FP_Linear_Form abs_error = FP_Linear_Form(error); + FP_Linear_Form result = linearized_first_operand + linearized_second_operand + + relative_error(linearized_first_operand) + + relative_error(linearized_second_operand) + + abs_error; + return result; +} + +} // namespace Parma_Polyhedra_Library + +#endif // !defined(PPL_Sum_Floating_Point_Expression_templates_hh)

1 0

[GIT] ppl/ppl(floating_point): Added a first implementation of the relative_error and intervalize
by Fabio Bossi 02 Sep '09

02 Sep '09

Module: ppl/ppl Branch: floating_point Commit: 67bec8827d8c0ce1db133ae9f20dd36830ffffe8 URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=67bec8827d8c… Author: Fabio Bossi <bossi(a)cs.unipr.it> Date: Wed Sep 2 14:41:51 2009 +0200 Added a first implementation of the relative_error and intervalize methods. Fixed a mistake with loops over linear forms. Several style improvements. --- src/Constant_Floating_Point_Expression.inlines.hh | 3 +- src/Floating_Point_Expression.defs.hh | 5 +- src/Floating_Point_Expression.inlines.hh | 6 +- src/Floating_Point_Expression.templates.hh | 87 +++++++++++++++++++++ src/Makefile.am | 1 + 5 files changed, 95 insertions(+), 7 deletions(-) diff --git a/src/Constant_Floating_Point_Expression.inlines.hh b/src/Constant_Floating_Point_Expression.inlines.hh index d1edd38..1fa15fb 100644 --- a/src/Constant_Floating_Point_Expression.inlines.hh +++ b/src/Constant_Floating_Point_Expression.inlines.hh @@ -34,6 +34,7 @@ Constant_Floating_Point_Expression(const boundary_type lb, const boundary_type ub) : value(lb) { assert(lb <= ub); + // FIXME: this may be incorrect for some policies. value.join_assign(ub); } @@ -54,10 +55,8 @@ inline typename Constant_Floating_Point_Expression<FP_Interval_Type, FP_Format>::FP_Linear_Form Constant_Floating_Point_Expression<FP_Interval_Type, FP_Format> ::linearize(const FP_Interval_Abstract_Store& store) const { - FP_Linear_Form result = FP_Linear_Form(value); return result; - } } // namespace Parma_Polyhedra_Library diff --git a/src/Floating_Point_Expression.defs.hh b/src/Floating_Point_Expression.defs.hh index 0d4e854..e712fd4 100644 --- a/src/Floating_Point_Expression.defs.hh +++ b/src/Floating_Point_Expression.defs.hh @@ -71,11 +71,11 @@ public: static bool overflows(const FP_Linear_Form& lf); + static FP_Linear_Form relative_error(const FP_Linear_Form&); + static FP_Linear_Form intervalize(const FP_Linear_Form&, const FP_Interval_Abstract_Store& store); - static FP_Linear_Form relative_error(const FP_Linear_Form&); - }; // class Floating_Point_Expression // Initialize static members of the class. @@ -90,5 +90,6 @@ Floating_Point_Expression<FP_Interval_Type, FP_Format>::absolute_error = } // namespace Parma_Polyhedra_Library #include "Floating_Point_Expression.inlines.hh" +#include "Floating_Point_Expression.templates.hh" #endif // !defined(PPL_Floating_Point_Expression_defs_hh) diff --git a/src/Floating_Point_Expression.inlines.hh b/src/Floating_Point_Expression.inlines.hh index 20f37e9..3e38dd2 100644 --- a/src/Floating_Point_Expression.inlines.hh +++ b/src/Floating_Point_Expression.inlines.hh @@ -36,12 +36,12 @@ template <typename FP_Interval_Type, typename FP_Format> inline bool Floating_Point_Expression<FP_Interval_Type, FP_Format> ::overflows(const FP_Linear_Form& lf) { - if(!lf.inhomogeneous_term().is_bounded()) + if (!lf.inhomogeneous_term().is_bounded()) return true; dimension_type dimension = lf.space_dimension(); - for(dimension_type i = 0; i <= dimension; ++i) { - if(!lf.coefficient(Variable(i)).is_bounded()) + for (dimension_type i = 0; i < dimension; ++i) { + if (!lf.coefficient(Variable(i)).is_bounded()) return true; } diff --git a/src/Floating_Point_Expression.templates.hh b/src/Floating_Point_Expression.templates.hh new file mode 100644 index 0000000..3ea080a --- /dev/null +++ b/src/Floating_Point_Expression.templates.hh @@ -0,0 +1,87 @@ +/* Floating_Point_Expression class implementation: + non-inline template functions. + Copyright (C) 2001-2009 Roberto Bagnara <bagnara(a)cs.unipr.it> + +This file is part of the Parma Polyhedra Library (PPL). + +The PPL 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 of the License, or (at your +option) any later version. + +The PPL 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. + +You should have received a copy of the GNU General Public License +along with this program; if not, write to the Free Software Foundation, +Inc., 51 Franklin Street, Fifth Floor, Boston, MA 02111-1307, USA. + +For the most up-to-date information see the Parma Polyhedra Library +site: http://www.cs.unipr.it/ppl/ . */ + +#ifndef PPL_Floating_Point_Expression_templates_hh +#define PPL_Floating_Point_Expression_templates_hh 1 + +#include "Floating_Point_Expression.defs.hh" +#include <cmath> + +namespace Parma_Polyhedra_Library { + +template<typename FP_Interval_Type, typename FP_Format> +typename Floating_Point_Expression<FP_Interval_Type, FP_Format>::FP_Linear_Form +Floating_Point_Expression<FP_Interval_Type, FP_Format> +::relative_error(const FP_Linear_Form& lf) { + /* FIXME: here we assume that boundary_type can represent + (2)^(-FP_Format::fraction_bits) precisely. */ + FP_Interval_Type error_propagator(pow(-2, -FP_Format::fraction_bits)); + // FIXME: this may be incorrect for some policies. + error_propagator.join_assign(FP_Interval_Type( + pow(2, -FP_Format::fraction_bits))); + + // Handle the inhomogeneous term. + FP_Interval_Type current_term = lf.inhomogeneous_term(); + boundary_type current_multiplier = std::max(abs(current_term.lower()), + abs(current_term.upper())); + FP_Linear_Form current_result_term = FP_Linear_Form(current_multiplier * + error_propagator); + FP_Linear_Form result = FP_Linear_Form(current_result_term); + + // Handle the other terms. + dimension_type dimension = lf.space_dimension(); + for (dimension_type i = 0; i < dimension; ++i) { + current_term = lf.coefficient(Variable(i)); + current_multiplier = std::max(abs(current_term.lower()), + abs(current_term.upper())); + current_result_term = FP_Linear_Form(Variable(i)) * current_multiplier * + error_propagator; + result += current_result_term; + } + + return result; +} + +template<typename FP_Interval_Type, typename FP_Format> +typename Floating_Point_Expression<FP_Interval_Type, FP_Format>::FP_Linear_Form +Floating_Point_Expression<FP_Interval_Type, FP_Format> +::intervalize(const FP_Linear_Form& lf, + const FP_Interval_Abstract_Store& store) { + FP_Interval_Type resulting_interval = lf.inhomogeneous_term(); + dimension_type dimension = lf.space_dimension(); + for (dimension_type i = 0; i < dimension; ++i) { + typename FP_Interval_Abstract_Store::const_iterator + next_variable_value = store.find(i); + if (next_variable_value != store.end()) { + FP_Interval_Type current_coefficient = lf.coefficient(Variable(i)); + resulting_interval += current_coefficient * (*next_variable_value); + } + } + + FP_Linear_Form result = FP_Interval_Type(resulting_interval); + return result; +} + +} // namespace Parma_Polyhedra_Library + +#endif // !defined(PPL_Floating_Point_Expression_templates_hh) diff --git a/src/Makefile.am b/src/Makefile.am index 89c9d1e..9fd2582 100644 --- a/src/Makefile.am +++ b/src/Makefile.am @@ -344,6 +344,7 @@ wrap_string.hh \ Floating_Point_Expression.types.hh \ Floating_Point_Expression.defs.hh \ Floating_Point_Expression.inlines.hh \ +Floating_Point_Expression.templates.hh \ Constant_Floating_Point_Expression.types.hh \ Constant_Floating_Point_Expression.defs.hh \ Constant_Floating_Point_Expression.inlines.hh \

1 0

[GIT] ppl/ppl(floating_point): Rewritten according to Enea's indications.
by Fabio Biselli 02 Sep '09

02 Sep '09

Module: ppl/ppl Branch: floating_point Commit: c8afa3a4e165f17b2cf71eae85770d4c300bc37c URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=c8afa3a4e165… Author: Fabio Biselli <fabio.biselli(a)studenti.unipr.it> Date: Wed Sep 2 13:34:01 2009 +0200 Rewritten according to Enea's indications. --- src/Difference_Floating_Point_Expression.defs.hh | 22 +++++++------ ...Difference_Floating_Point_Expression.inlines.hh | 35 +++++++++----------- src/Division_Floating_Point_Expression.defs.hh | 22 ++++++------ src/Division_Floating_Point_Expression.inlines.hh | 34 +++++++++---------- ...ultiplication_Floating_Point_Expression.defs.hh | 26 +++++++------- ...iplication_Floating_Point_Expression.inlines.hh | 35 ++++++++++--------- src/Sum_Floating_Point_Expression.defs.hh | 18 +++++----- src/Sum_Floating_Point_Expression.inlines.hh | 32 ++++++++--------- 8 files changed, 110 insertions(+), 114 deletions(-) Diff: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commitdiff;h=c8afa3a4…

1 0

[GIT] ppl/ppl(floating_point): Added swap functions and inhibited assignment and copy constructor.
by Fabio Biselli 02 Sep '09

02 Sep '09

Module: ppl/ppl Branch: floating_point Commit: b7ee345b53fd72a03d08d51295c72f97f71b7ab2 URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=b7ee345b53fd… Author: Fabio Biselli <fabio.biselli(a)studenti.unipr.it> Date: Tue Sep 1 20:06:03 2009 +0200 Added swap functions and inhibited assignment and copy constructor. --- src/Difference_Floating_Point_Expression.defs.hh | 39 ++++++++++++++--- ...Difference_Floating_Point_Expression.inlines.hh | 32 ++++++++++++-- src/Division_Floating_Point_Expression.defs.hh | 37 ++++++++++++++-- src/Division_Floating_Point_Expression.inlines.hh | 27 +++++++++++- ...ultiplication_Floating_Point_Expression.defs.hh | 44 ++++++++++++++++--- ...iplication_Floating_Point_Expression.inlines.hh | 46 +++++++++++++++----- src/Sum_Floating_Point_Expression.defs.hh | 40 ++++++++++++++--- src/Sum_Floating_Point_Expression.inlines.hh | 38 +++++++++++++--- 8 files changed, 254 insertions(+), 49 deletions(-) Diff: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commitdiff;h=b7ee345b…

1 0

[GIT] ppl/ppl(floating_point): Removed trailing whitespaces.
by Fabio Biselli 02 Sep '09

02 Sep '09

Module: ppl/ppl Branch: floating_point Commit: 625f2adc4b2dd4095a23203bdb64f9f58d4930aa URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=625f2adc4b2d… Author: Fabio Biselli <fabio.biselli(a)studenti.unipr.it> Date: Tue Sep 1 16:29:05 2009 +0200 Removed trailing whitespaces. --- src/Difference_Floating_Point_Expression.defs.hh | 4 ++-- ...Difference_Floating_Point_Expression.inlines.hh | 6 +++--- 2 files changed, 5 insertions(+), 5 deletions(-) diff --git a/src/Difference_Floating_Point_Expression.defs.hh b/src/Difference_Floating_Point_Expression.defs.hh index d38abb7..db8cf3b 100644 --- a/src/Difference_Floating_Point_Expression.defs.hh +++ b/src/Difference_Floating_Point_Expression.defs.hh @@ -1,4 +1,4 @@ -/* Declarations for the Difference_Floating_Point_Expression class and +/* Declarations for the Difference_Floating_Point_Expression class and its constituents. Copyright (C) 2001-2009 Roberto Bagnara <bagnara(a)cs.unipr.it> @@ -56,7 +56,7 @@ public: Floating_Point_Expression<FP_Interval_Type, FP_Format>* const num, Floating_Point_Expression<FP_Interval_Type, FP_Format>* const den); - ~Difference_Floating_Point_Expression(); + ~Difference_Floating_Point_Expression(); FP_Linear_Form linearize(const FP_Interval_Abstract_Store& store) const; diff --git a/src/Difference_Floating_Point_Expression.inlines.hh b/src/Difference_Floating_Point_Expression.inlines.hh index abd1559..b9a8c9e 100644 --- a/src/Difference_Floating_Point_Expression.inlines.hh +++ b/src/Difference_Floating_Point_Expression.inlines.hh @@ -1,4 +1,4 @@ -/* Difference_Floating_Point_Expression class implementation: inline +/* Difference_Floating_Point_Expression class implementation: inline functions. Copyright (C) 2001-2009 Roberto Bagnara <bagnara(a)cs.unipr.it> @@ -48,9 +48,9 @@ Difference_Floating_Point_Expression<FP_Interval_Type, FP_Format>:: delete second_operand; } - // FIXME: add function body + // FIXME: add function body template <typename FP_Interval_Type, typename FP_Format> -inline +inline typename Difference_Floating_Point_Expression<FP_Interval_Type, FP_Format>::FP_Linear_Form Difference_Floating_Point_Expression<FP_Interval_Type, FP_Format>::linearize(

1 0

[GIT] ppl/ppl(floating_point): Added a few more classes for floating point expressions.
by Fabio Biselli 02 Sep '09

02 Sep '09

Module: ppl/ppl Branch: floating_point Commit: 8f090b25bf4b632cb3832f74b02421dc3812c917 URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=8f090b25bf4b… Author: Fabio Biselli <fabio.biselli(a)studenti.unipr.it> Date: Tue Sep 1 16:08:08 2009 +0200 Added a few more classes for floating point expressions. --- src/Difference_Floating_Point_Expression.defs.hh | 75 ++++++++++++++++++++ ...Difference_Floating_Point_Expression.inlines.hh | 64 +++++++++++++++++ src/Difference_Floating_Point_Expression.types.hh | 22 ++++++ src/Division_Floating_Point_Expression.defs.hh | 75 ++++++++++++++++++++ src/Division_Floating_Point_Expression.inlines.hh | 63 ++++++++++++++++ src/Division_Floating_Point_Expression.types.hh | 22 ++++++ ...ultiplication_Floating_Point_Expression.defs.hh | 74 +++++++++++++++++++ ...iplication_Floating_Point_Expression.inlines.hh | 64 +++++++++++++++++ ...ltiplication_Floating_Point_Expression.types.hh | 22 ++++++ src/Sum_Floating_Point_Expression.defs.hh | 75 ++++++++++++++++++++ src/Sum_Floating_Point_Expression.inlines.hh | 64 +++++++++++++++++ src/Sum_Floating_Point_Expression.types.hh | 22 ++++++ 12 files changed, 642 insertions(+), 0 deletions(-) Diff: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commitdiff;h=8f090b25…

1 0

[GIT] ppl/ppl(floating_point): Corrected absolute error calculation to account for underflows.
by Fabio Bossi 02 Sep '09

02 Sep '09

Module: ppl/ppl Branch: floating_point Commit: d10726e73284885a6c609f6e6634e8aa1908d700 URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=d10726e73284… Author: Fabio Bossi <bossi(a)cs.unipr.it> Date: Wed Sep 2 11:23:28 2009 +0200 Corrected absolute error calculation to account for underflows. --- src/Floating_Point_Expression.defs.hh | 7 +++++-- 1 files changed, 5 insertions(+), 2 deletions(-) diff --git a/src/Floating_Point_Expression.defs.hh b/src/Floating_Point_Expression.defs.hh index f84c6a4..0d4e854 100644 --- a/src/Floating_Point_Expression.defs.hh +++ b/src/Floating_Point_Expression.defs.hh @@ -78,11 +78,14 @@ public: }; // class Floating_Point_Expression -// Initialize static members of the class +// Initialize static members of the class. template <typename FP_Interval_Type, typename FP_Format> typename Floating_Point_Expression<FP_Interval_Type, FP_Format>::boundary_type Floating_Point_Expression<FP_Interval_Type, FP_Format>::absolute_error = - pow(2, 1 - FP_Format::exponent_bias - FP_Format::fraction_bits); + std::max(pow(2, 1 - FP_Format::exponent_bias - FP_Format::fraction_bits), + std::numeric_limits<typename Floating_Point_Expression<FP_Interval_Type, + FP_Format> + ::boundary_type>::denorm_min()); } // namespace Parma_Polyhedra_Library

1 0

[GIT] ppl/ppl(floating_point): New commits
by Roberto Bagnara 02 Sep '09

02 Sep '09

URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=a2335bf94541… Author: Fabio Bossi <bossi(a)cs.unipr.it> Date: Wed Sep 2 11:06:44 2009 +0200 Added static helper methods and a few policy classes. URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=6c9d4e2385d9… Author: Fabio Bossi <bossi(a)cs.unipr.it> Date: Wed Sep 2 09:54:33 2009 +0200 Made constructors explicit. URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=675fae6564ba… Author: Fabio Bossi <bossi(a)cs.unipr.it> Date: Wed Sep 2 09:42:16 2009 +0200 Fixed code according to Enea's indications. URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=8b73f935478d… Author: Fabio Bossi <bossi(a)cs.unipr.it> Date: Tue Sep 1 18:05:12 2009 +0200 Added destructor implementation. URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=631b2220535f… Author: Fabio Bossi <bossi(a)cs.unipr.it> Date: Tue Sep 1 17:35:44 2009 +0200 Fixed a typo. URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=f9e4c9feaea4… Author: Fabio Bossi <bossi(a)cs.unipr.it> Date: Tue Sep 1 17:29:24 2009 +0200 Added swapping functions. URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=1f4b78b53ec1… Author: Fabio Bossi <bossi(a)cs.unipr.it> Date: Tue Sep 1 11:38:06 2009 +0200 Added the files for the first (incomplete) version of class Opposite_Floating_Point_Expression

1 0