Changeset 3688 for trunk/test_more

May 26, 2015 8:09:43 PM (5 years ago)

merge to branch: trunk
from repository:
start hash code: 1d42b78f58cf6afaf6d0274f2569aa586e7c5bd3
end hash code: ffed0cb6e1a7efb9ffcdc6296994d1793eb31b73

commit ffed0cb6e1a7efb9ffcdc6296994d1793eb31b73
Author: Brad Bell <bradbell@…>
Date: Tue May 26 16:33:27 2015 -0700

Squashed commit of the branch cond_exp_2:

commit 19267b3d325214190d5a8bac3011a8aa50326e09
Author: Brad Bell <bradbell@…>
Date: Tue May 26 16:33:00 2015 -0700

Automatically generated changes to auto-tools install.

commit 6040a4459bca0b26c03275893dbc52e27445862f
Author: Brad Bell <bradbell@…>
Date: Tue May 26 15:43:11 2015 -0700

base_require.hpp: add missin #, fix level of standard base heading.
whats_new_15.omh: minor improvement to base type mention.

commit 526a77f22c6544048c45f4094649dcd2b688b2ed
Author: Brad Bell <bradbell@…>
Date: Tue May 26 14:57:30 2015 -0700

div_op.hpp: Decided to keep zdouble(0.0) / 0.0 == 0.0 (too many changes).

commit dc8b28c5eadfc310db4ad8eca393c59c6f57a544
Author: Brad Bell <bradbell@…>
Date: Tue May 26 12:29:15 2015 -0700

div_op.hpp: prepare for no absolute zero on division op.
mul_zdouble.cpp: test all division and multiplication ops.

commit a020bf9d3216af6fb9fee30588b9edb56b03da71
Author: Brad Bell <bradbell@…>
Date: Tue May 26 10:20:23 2015 -0700

mul_zdouble.cpp: modify so easy to add other operators.

commit 04cff2e65aa9c133e8bf100e62c2c1d48de775cf
Author: Brad Bell <bradbell@…>
Date: Tue May 26 09:25:39 2015 -0700

Start comprehensive testing of zdouble with multi-level reverse.

base_require.hpp: document absolute zero, improve warning about API.
zdouble.hpp: improve nan and allow for cross reference.
whats_new_15.omh: mention change to base type requirements. base_require special case, remove trailing white space.
zdouble.cpp: some minor improvements.

commit 38c040a84f586133d77782579af0396f5e1b6f3c
Author: Brad Bell <bradbell@…>
Date: Tue May 26 06:26:15 2015 -0700

  1. Separate example discussion for source code using Source heading.
  2. Fix reference to double that should now be zdouble.

mul_level_adolc.cpp:a fix commnet w -> dw.
whats_new_15.omh: advance data when changes merged into master.

commit c4b42d45cc269278801039234e0d3225e3caba2f
Author: Brad Bell <bradbell@…>
Date: Tue May 26 05:59:42 2015 -0700

mul_level.omh: Imporve description of multi-level taping.

commit dde678cda228dc1fc0191fa0f49cd55cfdfa7d8a
Author: Brad Bell <bradbell@…>
Date: Tue May 26 05:12:10 2015 -0700

  1. Use or at least discuss, zdouble.
  2. Use a1 (a2) infront of variables with one (two) levels of AD.
  3. Use thread_alloc instead of omp_alloc (deprecated).

zdouble.hpp: make defintion of absolute zero a cross reference.
mul_level_adolc.cpp: delete trailing white space.
mul_level_adolc_ode.cpp: delete trailing white space.
mul_level.cpp: declare that an argument is const.

commit 8de5156761d33c1396e136283731e738a782a739
Author: Brad Bell <bradbell@…>
Date: Mon May 25 09:19:30 2015 -0700

zdouble.hpp: Specialize CppAD::nan so works with zdouble.
near_equal.hpp: fix zdouble case (nolonger need x==nan), remove trailing white space.
mul_level_ode.cpp: use zdouble and a1, a2 notation, remove trailing white space.
zdouble.cpp: test nan.
mul_level.omh: correct some typos.

commit bc60b4134799487a7dc6c6a6eb4d556c327e749c
Author: Brad Bell <bradbell@…>
Date: Sun May 24 14:29:25 2015 -0700

  1. Move change_const to change_param.

change_param.cpp: minor changes of names, remove trailing white space.
mul_level.cpp: Comment about when start taping at first level.
whats_new_12.omh: remove trialing white space.
whats_new_15.omh: remove trialing white space, comment on zdouble.

commit 0f66ac25e899a81173f23d4e846679951cf0cb0f
Author: Brad Bell <bradbell@…>
Date: Sun May 24 13:01:50 2015 -0700

optimize.hpp: remove call to deprecated epsilon function.
zdouble.hpp: fix section name.
mul_level.cpp: zdouble and more moder notation, remove trailing white space.
mul_level.omh: zdouble and more moder notation, remove trailing white space.

commit cd6b134dbc0a6a0e30ff570b4dc7e4b068853ffd
Author: Brad Bell <bradbell@…>
Date: Sun May 24 11:59:10 2015 -0700

zdouble.cpp: Motivating example (demonstrates problem and solution).

commit 8d7b8f92091588a5d6f63ca50395ce1dc4594133
Author: Brad Bell <bradbell@…>
Date: Sun May 24 10:40:29 2015 -0700

Add doxygen documentation for zdouble. expand macros to get operators defined by macros.
zdouble.hpp: add computed assignment from double.

commit 65761538a535e95e9683b15fd886306123bf6baa
Author: Brad Bell <bradbell@…>
Date: Sun May 24 07:44:52 2015 -0700

zdouble.hpp: Add user documentation for zdouble class.

commit 265fddb435a18abf2d0aa8a3d60bf373cb944eb3
Author: Brad Bell <bradbell@…>
Date: Sun May 24 06:18:52 2015 -0700

Define the class zdouble as a replacement for double. This is a
tentative solution to multi-level AD reverse mode propigating nans through
part of conditional expression that is not used (issue #8). automatically run doxygen if there is no doxygen.err. change to use zdouble instead of double.
user_ad.hpp: remove trailing white space.
zdouble.hpp: needs user and developer documentation and more testing.

1 added
4 edited


  • trunk/test_more/CMakeLists.txt

    r3684 r3688  
    112112        mul_eq.cpp
    113113        mul_level.cpp
     114        mul_zdouble.cpp
    114115        mul_zero_one.cpp
    115116        near_equal_ext.cpp
  • trunk/test_more/

    r3684 r3688  
    138138        mul_eq.cpp \
    139139        mul_level.cpp \
     140        mul_zdouble.cpp \
    140141        mul_zero_one.cpp \
    141142        near_equal_ext.cpp \
  • trunk/test_more/

    r3685 r3688  
    114114        forward_order.cpp from_base.cpp fun_check.cpp jacobian.cpp \
    115115        limits.cpp log.cpp log1p.cpp log10.cpp mul.cpp mul_eq.cpp \
    116         mul_level.cpp mul_zero_one.cpp near_equal_ext.cpp neg.cpp \
    117         ode_err_control.cpp optimize.cpp parameter.cpp poly.cpp \
    118         pow.cpp pow_int.cpp print_for.cpp romberg_one.cpp rosen_34.cpp \
    119         runge_45.cpp reverse.cpp rev_sparse_hes.cpp rev_sparse_jac.cpp \
    120         rev_two.cpp simple_vector.cpp sin.cpp sin_cos.cpp sinh.cpp \
     116        mul_level.cpp mul_zdouble.cpp mul_zero_one.cpp \
     117        near_equal_ext.cpp neg.cpp ode_err_control.cpp optimize.cpp \
     118        parameter.cpp poly.cpp pow.cpp pow_int.cpp print_for.cpp \
     119        romberg_one.cpp rosen_34.cpp runge_45.cpp reverse.cpp \
     120        rev_sparse_hes.cpp rev_sparse_jac.cpp rev_two.cpp \
     121        simple_vector.cpp sin.cpp sin_cos.cpp sinh.cpp \
    121122        sparse_hessian.cpp sparse_jacobian.cpp sparse_vec_ad.cpp \
    122123        sqrt.cpp std_math.cpp sub.cpp sub_eq.cpp sub_zero.cpp tan.cpp \
    144145        limits.$(OBJEXT) log.$(OBJEXT) log1p.$(OBJEXT) log10.$(OBJEXT) \
    145146        mul.$(OBJEXT) mul_eq.$(OBJEXT) mul_level.$(OBJEXT) \
    146         mul_zero_one.$(OBJEXT) near_equal_ext.$(OBJEXT) neg.$(OBJEXT) \
     147        mul_zdouble.$(OBJEXT) mul_zero_one.$(OBJEXT) \
     148        near_equal_ext.$(OBJEXT) neg.$(OBJEXT) \
    147149        ode_err_control.$(OBJEXT) optimize.$(OBJEXT) \
    148150        parameter.$(OBJEXT) poly.$(OBJEXT) pow.$(OBJEXT) \
    526528        mul_eq.cpp \
    527529        mul_level.cpp \
     530        mul_zdouble.cpp \
    528531        mul_zero_one.cpp \
    529532        near_equal_ext.cpp \
    672675@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/mul_eq.Po@am__quote@
    673676@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/mul_level.Po@am__quote@
     677@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/mul_zdouble.Po@am__quote@
    674678@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/mul_zero_one.Po@am__quote@
    675679@AMDEP_TRUE@@am__include@ @am__quote@./$(DEPDIR)/near_equal_ext.Po@am__quote@
  • trunk/test_more/test_more.cpp

    r3684 r3688  
    6868extern bool mul_level(void);
    6969extern bool MulEq(void);
     70extern bool mul_zdouble(void);
    7071extern bool MulZeroOne(void);
    7172extern bool NearEqualExt(void);
    183184        ok &= Run( mul_level,       "mul_level"      );
    184185        ok &= Run( MulEq,           "MulEq"          );
     186        ok &= Run( mul_zdouble,     "mul_zdouble"    );
    185187        ok &= Run( MulZeroOne,      "MulZeroOne"     );
    186188        ok &= Run( NearEqualExt,    "NearEqualExt"   );
Note: See TracChangeset for help on using the changeset viewer.