September 2009 - PPL-devel - BUGSENG Mailman

[GIT] ppl/ppl(pip): Added additional check in simplex, leading to simpler decision trees.
by François Galea 24 Sep '09

24 Sep '09

Module: ppl/ppl Branch: pip Commit: a18c563be178e4e00a32720c6e2ce236a3cf7a81 URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=a18c563be178… Author: François Galea <francois.galea(a)uvsq.fr> Date: Thu Sep 24 11:55:41 2009 +0200 Added additional check in simplex, leading to simpler decision trees. --- src/PIP_Tree.cc | 33 +++++++++++++++++++++++++++++++++ tests/PIP_Problem/pipproblem.cc | 11 +++++++---- 2 files changed, 40 insertions(+), 4 deletions(-) diff --git a/src/PIP_Tree.cc b/src/PIP_Tree.cc index 1f904eb..1012838 100644 --- a/src/PIP_Tree.cc +++ b/src/PIP_Tree.cc @@ -713,6 +713,39 @@ PIP_Solution_Node::solve(PIP_Tree_Node*& parent_ref, } } + /* If there remains a row i with undetermined sign and at least one + positive S_ij coefficient, where constraint t_i(z) > 0 is not + compatible with the context, the row parameter can be considered + negative + */ + if (i_ == n_a_d && i__ != n_a_d) { + for (i=i__; i<num_rows; ++i) { + if (sign[i] != MIXED) + continue; + bool found = false; + const Row &p = tableau.s[i]; + for (j=0; j<num_vars; ++j) + if (p[j] > 0) { + found = true; + break; + } + if (!found) + continue; + Row row(tableau.t[i]); + row[0] -= 1; + if (compatibility_check(context, row)) { + if (i__ == n_a_d) + i__ = i; + } else { + sign[i] = NEGATIVE; + if (i_ == n_a_d) + i_ = i; + if (i__ == i) + i__ = n_a_d; + } + } + } + /* If we have found a row i_ with negative parameters : Either the problem is empty, or a pivoting step is required */ diff --git a/tests/PIP_Problem/pipproblem.cc b/tests/PIP_Problem/pipproblem.cc index 8d84a95..6050cd0 100644 --- a/tests/PIP_Problem/pipproblem.cc +++ b/tests/PIP_Problem/pipproblem.cc @@ -32,7 +32,8 @@ display_solution(const PIP_Tree pip, const Variables_Set &vars, int indent=0) { nout << setw(indent*2) << "" << "_|_" << endl; } else { const Constraint_System &constraints = pip->constraints(); - if (!constraints.empty()) { + bool constraints_empty = constraints.empty(); + if (!constraints_empty) { nout << setw(indent*2) << "" << "if "; Constraint_System::const_iterator begin = constraints.begin(); Constraint_System::const_iterator end = constraints.end(); @@ -51,12 +52,14 @@ display_solution(const PIP_Tree pip, const Variables_Set &vars, int indent=0) { Variables_Set::const_iterator begin = vars.begin(); Variables_Set::const_iterator end = vars.end(); Variables_Set::const_iterator i; - nout << setw(indent*2+2) << "" << "{"; + nout << setw(indent*2+(constraints_empty?0:2)) << "" << "{"; for (i=begin; i!=end; ++i) nout << ((i==begin)?"":" ; ") << sn->parametric_values(Variable(*i)); nout << "}" << endl; - nout << setw(indent*2) << "" << "else" << endl; - nout << setw(indent*2+2) << "" << "_|_" << endl; + if (!constraints_empty) { + nout << setw(indent*2) << "" << "else" << endl; + nout << setw(indent*2+2) << "" << "_|_" << endl; + } } } }

1 0

[GIT] ppl/ppl(floating_point): Completed the method refine_with_linear_form_inequality();
by Fabio Biselli 24 Sep '09

24 Sep '09

Module: ppl/ppl Branch: floating_point Commit: a21e78b3e9bb55c5ad53ad4feb06b7640e7230c5 URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=a21e78b3e9bb… Author: Fabio Biselli <fabio.biselli(a)studenti.unipr.it> Date: Thu Sep 24 14:09:22 2009 +0200 Completed the method refine_with_linear_form_inequality(); added the auxiliary function left_two_var_refine(). --- src/BD_Shape.defs.hh | 10 ++ src/BD_Shape.templates.hh | 252 ++++++++++++++++++++------------------------- 2 files changed, 123 insertions(+), 139 deletions(-) Diff: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commitdiff;h=a21e78b3…

1 0

[GIT] ppl/ppl(floating_point): Always assert that all variables in the abstract store are bounded.
by Fabio Bossi 24 Sep '09

24 Sep '09

Module: ppl/ppl Branch: floating_point Commit: 3ad560b33c28a63c68d5f0928275ae3b28255bf0 URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=3ad560b33c28… Author: Fabio Bossi <bossi(a)cs.unipr.it> Date: Thu Sep 24 11:09:02 2009 +0200 Always assert that all variables in the abstract store are bounded. Corrected the documentation accordingly. --- src/Polyhedron.defs.hh | 7 +++++-- src/Polyhedron.templates.hh | 3 +++ 2 files changed, 8 insertions(+), 2 deletions(-) diff --git a/src/Polyhedron.defs.hh b/src/Polyhedron.defs.hh index f2854b4..41c8ae6 100644 --- a/src/Polyhedron.defs.hh +++ b/src/Polyhedron.defs.hh @@ -995,7 +995,7 @@ public: \param store The interval abstract store in which the test defining the filter is performed. All variables that occur in \p left or \p right MUST have - an associated value in it. + an associated interval in it, and all of these intervals MUST be bounded. \exception std::invalid_argument Thrown if \p left (or \p right) is dimension-incompatible with \p *this. @@ -1201,7 +1201,8 @@ public: \param store The interval abstract store in which the variable assignment is performed. - All variables that occur in \p lf MUST have an associated value in it. + All variables that occur in \p lf MUST have an associated interval in it, + and all of these intervals MUST be bounded. \exception std::invalid_argument Thrown if \p lf and \p *this are dimension-incompatible or if \p var is @@ -2647,6 +2648,8 @@ protected: \param store The interval abstract store in which the approximation is performed. + All variables that occur in \p lf MUST have an associated interval in it, + and all of these intervals MUST be bounded. \param result Used to store the result. diff --git a/src/Polyhedron.templates.hh b/src/Polyhedron.templates.hh index 0ef06d2..491bb20 100644 --- a/src/Polyhedron.templates.hh +++ b/src/Polyhedron.templates.hh @@ -307,6 +307,7 @@ Polyhedron::refine_with_linear_form_inequality( "Polyhedron::refine_with_linear_form_inequality:" " FP_Format not a floating point type."); + PPL_ASSERT(store.is_bounded()); // Dimension compatibility checks. // The dimensions of left and right should not be greater than the // dimension of *this. @@ -352,6 +353,7 @@ const Box< Interval<FP_Format, Interval_Info> >& store) { "Polyhedron::affine_image:" " FP_Format not a floating point type."); + PPL_ASSERT(store.is_bounded()); // Dimension compatibility checks. // The dimension of lf should not be greater than the dimension of *this. const dimension_type lf_space_dim = lf.space_dimension(); @@ -398,6 +400,7 @@ Polyhedron::overapproximate_linear_form( " FP_Format not a floating point type."); PPL_ASSERT(lf_dimension <= store.space_dimension()); + PPL_ASSERT(store.is_bounded()); typedef Interval<FP_Format, Interval_Info> FP_Interval_Type; typedef Linear_Form<FP_Interval_Type> FP_Linear_Form;

1 0

[GIT] ppl/ppl(floating_point): Started to test Polyhedron:: affine_image on FP_Linear_Form.
by Roberto Amadini 24 Sep '09

24 Sep '09

Module: ppl/ppl Branch: floating_point Commit: 4c93dcb4734f0fc815eeed69bcb4fbc90f1a50f1 URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=4c93dcb4734f… Author: Roberto Amadini <r.amadini(a)virgilio.it> Date: Thu Sep 24 10:38:03 2009 +0200 Started to test Polyhedron::affine_image on FP_Linear_Form. --- src/Polyhedron.templates.hh | 2 +- tests/Floating_Point_Expression/Makefile.am | 23 +++++++++++++---------- 2 files changed, 14 insertions(+), 11 deletions(-) diff --git a/src/Polyhedron.templates.hh b/src/Polyhedron.templates.hh index 880d88b..0ef06d2 100644 --- a/src/Polyhedron.templates.hh +++ b/src/Polyhedron.templates.hh @@ -414,7 +414,7 @@ Polyhedron::overapproximate_linear_form( FP_Format curr_lb = curr_coeff.lower(); FP_Format curr_ub = curr_coeff.upper(); if (curr_lb != 0 || curr_ub != 0) { - const FP_Interval_Type& curr_int = store.get_iterval(Variable(i)); + const FP_Interval_Type& curr_int = store.get_interval(Variable(i)); FP_Interval_Type curr_addend(curr_ub - curr_lb); curr_addend *= aux_divisor2; curr_addend *= curr_int; diff --git a/tests/Floating_Point_Expression/Makefile.am b/tests/Floating_Point_Expression/Makefile.am index dc2e993..6698b36 100644 --- a/tests/Floating_Point_Expression/Makefile.am +++ b/tests/Floating_Point_Expression/Makefile.am @@ -51,11 +51,12 @@ $(top_builddir)/src/libppl.la \ @extra_libraries@ ORIGINAL_TESTS = \ -bdshape1 \ -floatingpointexpr1 \ -linearform1 \ -octagonalshape1 \ -octagonalshape2 +polyhedron1 +#bdshape1 \ +#floatingpointexpr1 \ +#linearform1 \ +#octagonalshape1 \ +#octagonalshape2 DERIVED_TESTS = @@ -125,15 +126,17 @@ print_INSTANCES: # Sources for the tests # -bdshape1_SOURCES = bdshape1.cc +polyhedron1_SOURCES = polyhedron1.cc -floatingpointexpr1_SOURCES = floatingpointexpr1.cc +#bdshape1_SOURCES = bdshape1.cc -linearform1_SOURCES = linearform1.cc +#floatingpointexpr1_SOURCES = floatingpointexpr1.cc -octagonalshape1_SOURCES = octagonalshape1.cc +#linearform1_SOURCES = linearform1.cc -octagonalshape2_SOURCES = octagonalshape2.cc +#octagonalshape1_SOURCES = octagonalshape1.cc + +#octagonalshape2_SOURCES = octagonalshape2.cc #

1 0

[GIT] ppl/ppl(floating_point): Removed commented code in floatingpointexpr1 .cc
by Roberto Amadini 23 Sep '09

23 Sep '09

Module: ppl/ppl Branch: floating_point Commit: 7ac2752fde0d8085804bd97bd3f2de4a76ca602e URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=7ac2752fde0d… Author: Roberto Amadini <r.amadini(a)virgilio.it> Date: Wed Sep 23 21:30:33 2009 +0200 Removed commented code in floatingpointexpr1.cc --- .../floatingpointexpr1.cc | 43 +------------------ 1 files changed, 3 insertions(+), 40 deletions(-) diff --git a/tests/Floating_Point_Expression/floatingpointexpr1.cc b/tests/Floating_Point_Expression/floatingpointexpr1.cc index 25b7b14..c4322eb 100644 --- a/tests/Floating_Point_Expression/floatingpointexpr1.cc +++ b/tests/Floating_Point_Expression/floatingpointexpr1.cc @@ -65,27 +65,7 @@ test02() { << known_result << endl; bool ok = (result == known_result); - return ok; /* - FP_Interval_Abstract_Store store_db(2); - store_db.set_interval(Variable(0), FP_Interval(0)); - store_db.set_interval(Variable(1), FP_Interval(4)); - Con_FP_Expression* con_db = new Con_FP_Expression(5, 6); - Var_FP_Expression* var0_db = new Var_FP_Expression(0); - Var_FP_Expression* var1_db = new Var_FP_Expression(1); - Sum_FP_Expression* sum_db = new Sum_FP_Expression(con_db, var1_db); - Mul_FP_Expression mul_db(var0_db, sum_db); - FP_Linear_Form result_db; - mul_db.linearize(store_db, FP_Linear_Form_Abstract_Store(), result_db); - - FP_Interval kr_db(-FP_Expression::absolute_error); - kr_db.join_assign(FP_Expression::absolute_error); - FP_Linear_Form known_result_db(kr_db); - - nout << "*** known_result_db ***" << endl - << known_result_db << endl; - bool ok_db = (result_db == known_result_db); - - return ok_fl && ok_db; */ + return ok; } // Tests linearization of variables in a given linear form abstract store. @@ -103,20 +83,7 @@ test03() { << known_result << endl; bool ok = (result == known_result); - return ok; /* - FP_Linear_Form_Abstract_Store store_db; - FP_Linear_Form known_result_db = - FP_Linear_Form(A); - store_db[0] = known_result_db; - Var_FP_Expression var_db(0); - FP_Linear_Form result_db; - var_db.linearize(FP_Interval_Abstract_Store(0), store_db, result_db); - - nout << "*** known_result_db ***" << endl - << known_result_db << endl; - bool ok_db = (result_db == known_result_db); - - return ok_fl && ok_db; */ + return ok; } // Tests the linearization of A + B. @@ -303,11 +270,7 @@ test09() { nout << "*** known_result1 ***" << endl << known_result1 << endl; bool ok1 = (known_result1 == result); -/* - ANALYZER_FP_FORMAT max_db = std::numeric_limits<ANALYZER_FP_FORMAT>::max(); - FP_Interval min_db = FP_Interval( - -FP_Expression::absolute_error); - min_db.join_assign(FP_Expression::absolute_error); */ + FP_Linear_Form known_result2 = FP_Linear_Form(min); Con_FP_Expression* con4 = new Con_FP_Expression(0, 0); Con_FP_Expression* con5 = new Con_FP_Expression(0, max);

1 0

[GIT] ppl/ppl(floating_point): Improved tests. Now they all pass and reach the maximum coverage.
by Roberto Amadini 23 Sep '09

23 Sep '09

Module: ppl/ppl Branch: floating_point Commit: 102eb141ca6f56abb3b38bcfb9d724f1f6c8f7c8 URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=102eb141ca6f… Author: Roberto Amadini <r.amadini(a)virgilio.it> Date: Wed Sep 23 21:23:38 2009 +0200 Improved tests. Now they all pass and reach the maximum coverage. --- .../floatingpointexpr1.cc | 85 ++++++++++---------- tests/Floating_Point_Expression/octagonalshape2.cc | 7 +- 2 files changed, 43 insertions(+), 49 deletions(-) diff --git a/tests/Floating_Point_Expression/floatingpointexpr1.cc b/tests/Floating_Point_Expression/floatingpointexpr1.cc index 63c978a..25b7b14 100644 --- a/tests/Floating_Point_Expression/floatingpointexpr1.cc +++ b/tests/Floating_Point_Expression/floatingpointexpr1.cc @@ -47,24 +47,25 @@ test01() { // Tests multiplication by zero. bool test02() { - FP_Interval_Abstract_Store store_fl(2); - store_fl.set_interval(Variable(0), FP_Interval(0)); - store_fl.set_interval(Variable(1), FP_Interval(10)); - Con_FP_Expression* con_fl = new Con_FP_Expression(5, 6); - Var_FP_Expression* var0_fl = new Var_FP_Expression(0); - Var_FP_Expression* var1_fl = new Var_FP_Expression(1); - Dif_FP_Expression* dif_fl = new Dif_FP_Expression(var1_fl, con_fl); - Mul_FP_Expression mul_fl(dif_fl, var0_fl); - FP_Linear_Form result_fl; - mul_fl.linearize(store_fl, FP_Linear_Form_Abstract_Store(), result_fl); - FP_Interval kr_fl(-FP_Expression::absolute_error); - kr_fl.join_assign(FP_Expression::absolute_error); - FP_Linear_Form known_result_fl(kr_fl); - - nout << "*** known_result_fl ***" << endl - << known_result_fl << endl; - bool ok_fl = (result_fl == known_result_fl); + FP_Interval_Abstract_Store store(2); + store.set_interval(Variable(0), FP_Interval(0)); + store.set_interval(Variable(1), FP_Interval(10)); + Con_FP_Expression* con = new Con_FP_Expression(5, 6); + Var_FP_Expression* var0 = new Var_FP_Expression(0); + Var_FP_Expression* var1 = new Var_FP_Expression(1); + Dif_FP_Expression* dif = new Dif_FP_Expression(var1, con); + Mul_FP_Expression mul(dif, var0); + FP_Linear_Form result; + mul.linearize(store, FP_Linear_Form_Abstract_Store(), result); + FP_Interval kr(-FP_Expression::absolute_error); + kr.join_assign(FP_Expression::absolute_error); + FP_Linear_Form known_result(kr); + + nout << "*** known_result ***" << endl + << known_result << endl; + bool ok = (result == known_result); + return ok; /* FP_Interval_Abstract_Store store_db(2); store_db.set_interval(Variable(0), FP_Interval(0)); store_db.set_interval(Variable(1), FP_Interval(4)); @@ -84,24 +85,25 @@ test02() { << known_result_db << endl; bool ok_db = (result_db == known_result_db); - return ok_fl && ok_db; + return ok_fl && ok_db; */ } // Tests linearization of variables in a given linear form abstract store. bool test03() { - FP_Linear_Form_Abstract_Store store_fl; + FP_Linear_Form_Abstract_Store store; Variable A(0); - FP_Linear_Form known_result_fl = FP_Linear_Form(A); - store_fl[0] = known_result_fl; - Var_FP_Expression var_fl(0); - FP_Linear_Form result_fl; - var_fl.linearize(FP_Interval_Abstract_Store(0), store_fl, result_fl); + FP_Linear_Form known_result = FP_Linear_Form(A); + store[0] = known_result; + Var_FP_Expression var(0); + FP_Linear_Form result; + var.linearize(FP_Interval_Abstract_Store(0), store, result); - nout << "*** known_result_fl ***" << endl - << known_result_fl << endl; - bool ok_fl = (result_fl == known_result_fl); + nout << "*** known_result ***" << endl + << known_result << endl; + bool ok = (result == known_result); + return ok; /* FP_Linear_Form_Abstract_Store store_db; FP_Linear_Form known_result_db = FP_Linear_Form(A); @@ -114,7 +116,7 @@ test03() { << known_result_db << endl; bool ok_db = (result_db == known_result_db); - return ok_fl && ok_db; + return ok_fl && ok_db; */ } // Tests the linearization of A + B. @@ -129,22 +131,18 @@ test04() { Sum_FP_Expression sum(var0, var1); FP_Linear_Form result; sum.linearize(store, FP_Linear_Form_Abstract_Store(), result); - nout << result << endl; Variable A(0); Variable B(1); FP_Linear_Form known_result = FP_Linear_Form(A); ANALYZER_FP_FORMAT exp = pow(2, -static_cast<ANALYZER_FP_FORMAT>(ANALYZED_FP_FORMAT::MANTISSA_BITS)); - nout << exp << endl; tmp = FP_Interval(1); tmp -= exp; FP_Interval tmp2(1); tmp2 += exp; tmp.join_assign(tmp2); - nout << tmp << endl; known_result *= tmp; known_result += tmp * FP_Linear_Form(B); - nout << FP_Expression::absolute_error << endl; tmp = FP_Interval(-FP_Expression::absolute_error); tmp.join_assign(FP_Expression::absolute_error); known_result += tmp; @@ -289,13 +287,12 @@ test08() { // Tests linearization of multiplication by unbounded operands. bool test09() { - ANALYZER_FP_FORMAT max_fl = std::numeric_limits<ANALYZER_FP_FORMAT>::max(); - FP_Interval min_fl = FP_Interval( - -FP_Expression::absolute_error); - min_fl.join_assign(FP_Expression::absolute_error); - FP_Linear_Form known_result1 = FP_Linear_Form(min_fl); + ANALYZER_FP_FORMAT max = std::numeric_limits<ANALYZER_FP_FORMAT>::max(); + FP_Interval min = FP_Interval(-FP_Expression::absolute_error); + min.join_assign(FP_Expression::absolute_error); + FP_Linear_Form known_result1 = FP_Linear_Form(min); Con_FP_Expression* con1 = new Con_FP_Expression(0, 0); - Con_FP_Expression* con2 = new Con_FP_Expression(0, max_fl); + Con_FP_Expression* con2 = new Con_FP_Expression(0, max); Sum_FP_Expression* sum = new Sum_FP_Expression(con1, con2); Con_FP_Expression* con3 = new Con_FP_Expression(0, 0); Mul_FP_Expression mul(con3, sum); @@ -306,14 +303,14 @@ test09() { nout << "*** known_result1 ***" << endl << known_result1 << endl; bool ok1 = (known_result1 == result); - +/* ANALYZER_FP_FORMAT max_db = std::numeric_limits<ANALYZER_FP_FORMAT>::max(); FP_Interval min_db = FP_Interval( -FP_Expression::absolute_error); - min_db.join_assign(FP_Expression::absolute_error); - FP_Linear_Form known_result2 = FP_Linear_Form(min_db); + min_db.join_assign(FP_Expression::absolute_error); */ + FP_Linear_Form known_result2 = FP_Linear_Form(min); Con_FP_Expression* con4 = new Con_FP_Expression(0, 0); - Con_FP_Expression* con5 = new Con_FP_Expression(0, max_db); + Con_FP_Expression* con5 = new Con_FP_Expression(0, max); Sum_FP_Expression* sum2 = new Sum_FP_Expression(con4, con5); Con_FP_Expression* con6 = new Con_FP_Expression(0, 0); Mul_FP_Expression mul2(sum2, con6); @@ -326,10 +323,10 @@ test09() { bool ok2 = (known_result2 == result2); Con_FP_Expression* con7 = new Con_FP_Expression(0, 0); - Con_FP_Expression* con8 = new Con_FP_Expression(0, max_db); + Con_FP_Expression* con8 = new Con_FP_Expression(0, max); Sum_FP_Expression* sum3 = new Sum_FP_Expression(con7, con8); Con_FP_Expression* con9 = new Con_FP_Expression(0, 0); - Con_FP_Expression* con10 = new Con_FP_Expression(0, max_db); + Con_FP_Expression* con10 = new Con_FP_Expression(0, max); Sum_FP_Expression* sum4 = new Sum_FP_Expression(con9, con10); Mul_FP_Expression mul3(sum3, sum4); diff --git a/tests/Floating_Point_Expression/octagonalshape2.cc b/tests/Floating_Point_Expression/octagonalshape2.cc index 0111f23..5a869ec 100644 --- a/tests/Floating_Point_Expression/octagonalshape2.cc +++ b/tests/Floating_Point_Expression/octagonalshape2.cc @@ -404,10 +404,8 @@ test10() { } // tests Octagonal_Shape<T>::refine_fp_interval_abstract_store -// FIXME: this test should be parametric according to the floating point -// format of analyzer and analyzed. bool -test11() { /* +test11() { Variable A(0); Variable B(1); FP_Interval_Abstract_Store store(2); @@ -446,8 +444,7 @@ test11() { /* bool ok2 = (tmp == known_result2); - return ok1 && ok2; */ - return true; + return ok1 && ok2; } } //namespace

1 0

[GIT] ppl/ppl(floating_point): Partial implementation of refine_with_linear_form_inequality();
by Fabio Biselli 23 Sep '09

23 Sep '09

Module: ppl/ppl Branch: floating_point Commit: de2be3f1b7f7b57f85b597a96e7373b169aac181 URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=de2be3f1b7f7… Author: Fabio Biselli <fabio.biselli(a)studenti.unipr.it> Date: Wed Sep 23 20:19:49 2009 +0200 Partial implementation of refine_with_linear_form_inequality(); Two auxiliary function added. --- src/BD_Shape.defs.hh | 49 ++++++ src/BD_Shape.templates.hh | 383 +++++++++++++++++++++++++++++++++++++++++++++ 2 files changed, 432 insertions(+), 0 deletions(-) Diff: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commitdiff;h=de2be3f1…

1 0

[GIT] ppl/ppl(floating_point): Fixed indentation.
by Roberto Amadini 23 Sep '09

23 Sep '09

Module: ppl/ppl Branch: floating_point Commit: f762ecded38d87c9ea2ab50d0f8dea0149f8f98b URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=f762ecded38d… Author: Roberto Amadini <r.amadini(a)virgilio.it> Date: Wed Sep 23 18:19:17 2009 +0200 Fixed indentation. --- .../floatingpointexpr1.cc | 3 ++- 1 files changed, 2 insertions(+), 1 deletions(-) diff --git a/tests/Floating_Point_Expression/floatingpointexpr1.cc b/tests/Floating_Point_Expression/floatingpointexpr1.cc index a29b649..63c978a 100644 --- a/tests/Floating_Point_Expression/floatingpointexpr1.cc +++ b/tests/Floating_Point_Expression/floatingpointexpr1.cc @@ -204,7 +204,8 @@ test06() { tmp = FP_Interval(-FP_Expression::absolute_error); tmp.join_assign(FP_Expression::absolute_error); - ANALYZER_FP_FORMAT exp = pow(2,-static_cast<ANALYZER_FP_FORMAT>((ANALYZED_FP_FORMAT::MANTISSA_BITS-1))); + ANALYZER_FP_FORMAT exp = pow(2, + -static_cast<ANALYZER_FP_FORMAT>((ANALYZED_FP_FORMAT::MANTISSA_BITS-1))); FP_Interval coeff = FP_Interval(2); coeff -= exp; FP_Interval coeff2(2);

1 0

[GIT] ppl/ppl(floating_point): Parametrized tests for Octagonal_Shape.
by Fabio Bossi 23 Sep '09

23 Sep '09

Module: ppl/ppl Branch: floating_point Commit: 8dd84af2cadee80bd897c0e0eb7822e09b4b9193 URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=8dd84af2cade… Author: Fabio Bossi <bossi(a)cs.unipr.it> Date: Wed Sep 23 18:16:21 2009 +0200 Parametrized tests for Octagonal_Shape. --- tests/Floating_Point_Expression/octagonalshape1.cc | 80 ++++++++-------- tests/Floating_Point_Expression/octagonalshape2.cc | 102 ++++++++++---------- 2 files changed, 91 insertions(+), 91 deletions(-) Diff: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commitdiff;h=8dd84af2…

1 0

[GIT] ppl/ppl(floating_point): Parametrized linearform1.cc and bdshape1.cc
by Roberto Amadini 23 Sep '09

23 Sep '09

Module: ppl/ppl Branch: floating_point Commit: f47bdef6918e15edbfb52084eead85941d563316 URL: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commit;h=f47bdef6918e… Author: Roberto Amadini <r.amadini(a)virgilio.it> Date: Wed Sep 23 18:11:37 2009 +0200 Parametrized linearform1.cc and bdshape1.cc Fixed indentation and removed an useless test in floatingpointexpr1.cc --- tests/Floating_Point_Expression/bdshape1.cc | 80 +++++++------- .../floatingpointexpr1.cc | 60 +++-------- tests/Floating_Point_Expression/linearform1.cc | 110 ++++++++++---------- 3 files changed, 112 insertions(+), 138 deletions(-) Diff: http://www.cs.unipr.it/git/gitweb.cgi?p=ppl/ppl.git;a=commitdiff;h=f47bdef6…

1 0