MetaPhysicL/NEWS at master · libMesh/MetaPhysicL · GitHub

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
1.4.0 -> 2.0.0
  * Improve nvc++ compatibility
  * Improved parallel testing
  * Enable and address paranoid warnings
  * Eliminate code deprecated in c++17
  * Update autotools
  * Avoid and test for spurious floating point exceptions
  * Remove pessimizing std::move
  * Update autoconf submodule
  * vexcl.m4 fix when only part of boost is installed
  * Allow dual numbers to execute on device
    * Focuses on the SemiDynamicSparseNumberArrayGeneric derivative container
    * Public do_derivatives static member deprecated
    * Float and long double overloads removed
    * Introduction of generic DynamicArrayWrapper
      * Template using declarations for wrapping std::array and KokkosArray
      * N template parameter changed from a typename to size_t
      * typename N parameter moved to SemiDynamicSparseNumberArrayGeneric since
        base template DynamicSparseNumberBase requires parameter pack of typename

1.3.3 -> 1.4.0
  * Throw exception in opt mode if DynamicStdArrayWrapper size exceeded
  * Add tolerance argument to DynamicSparseNumberBase::sparsity_trim()
  * ReplaceAlgebraicType metafunction
  * ValueType overloads for std::array, std::vector
  * MPI reduce support for DualNumber, SemiDynamicNumberArray, NumberArray
  * More flexible assertion macros
  * std::isinf overloads for DualNumber, DualExpression
  * Unit testing with FP exceptions enabled
  * Improved binary DualNumber operation speed + FPE safety
  * Assorted bug fixes, compiler workarounds, unit tests, unit test fixes

1.3.2 -> 1.3.3
  * TIMPI configuration bugfix

1.3.1 -> 1.3.2
  * Unit test options for disabling long double tests
  * Use SemiPermanent for derived StandardTypes
  * Fix assertions in wrapper constructor/assignment
  * --with-timpi-method configure option

1.3.0 -> 1.3.1
  * Build system fixes for "make check/distcheck" in some configurations

1.2.2 -> 1.3.0
  * StandardType specialization for SemiDynamicSparseNumberArray
  * StandardType specialization for DynamicStdArrayWrapper
  * Add template boolean parameter for StandardType and Packing of DualNumber

1.2.1 -> 1.2.2
  * Restrict implicit conversions (fix for Eigen compatibility)

1.2.0 -> 1.2.1
  * New configure options for selecting C++ standard

1.1.0 -> 1.2.0
  * C++17 standard library compatibility

1.0.2 -> 1.1.0
  * Add RawType<vector<DualNumber>> and RawType<vector<vector<DualNumber>>>
  * Add IsRawSame helper
  * Use new TIMPI NotADataType class
  * Use "-f" when using mv to overwrite our own files

1.0.0 -> 1.0.2
  * Autoconf submodule refactoring

0.6.0 -> 1.0.0
  * MetaPhysicL may now be optionally configured with TIMPI
  * Add StandardType and Packing specializations for parallel communication of
    DualNumber and DynamicSparseNumberArray data
  * DualNumbers can now have derivative initialization/calculation turned off.
    Adds a boolean template argument to DualNumber

0.5.2 -> 0.6.0
  * Implement SemiDynamicSparseNumberArray which leverages sparse functionality
    like DynamicSparseNumberArray/Vector but which uses static std::array
    background storage
  * Optimization for DualNumber product and quotient rules

0.5.1 -> 0.5.2
  * Defined CompareTypes between T and std::complex<T>

0.5.0 -> 0.5.1
  * Optimizations for multiply and derivative operations of DualNumbers using
    statically sized derivative storage types
  * Added robustness for pow(Dual*)

0.4.0 -> 0.5.0
  * Create metaphysicl_dbg_var to eliminate compiler warnings
  * Add unsigned long / long long support
  * Add long int type support
  * Fix Perl GetOptions unknown option error for mingw64
  * Fix warning with clang 7.x: Prevent copying when returning by name; use std::move

0.3.3 -> 0.4.0
  * Fix norm derivative for DualNumber
  * Add some assignment overloads for DualNumber
  * Don't default initialize NumberArray data
  * Test calculations of DualNumber<T, NumberArray<N,T>>
  * Add support for many C++11 math functions

0.3.2 -> 0.3.3
  * Begin incrementing library versioning properly
  * More declarations properly added to _decl headers
  * Fixes for numerous compiler warnings

0.3.1 -> 0.3.2
  * Support for std::complex functions for some classes
  * Bugfix in DynamicSparseNumber*::operator[]
  * More unit tests

0.3.0 -> 0.3.1
  * Better C++11/C++14 manual+auto configuration
  * const-compatible DualNumberSurrogate copy constructor/assignment
  * DualNumber construction/assignment from surrogates
  * NotADuckDualNumber derivatives fixes

0.2.0 -> 0.3.0
  * Build system fixes
  * Compilation fix on newer compilers
  * DualNumberSurrogate, DualNumber-wrapping-array capability

0.1.0 -> 0.2.0
  * Build system fixes
  * New unit test
  * Many more assorted bug fixes

0.0.0 -> 0.1.0
  * Compile-time sparse array/vector types
  * Run-time sparse array/vector types
  * DualExpression for performance with expression templates
  * NamedIndexArray for physics kernel experimentation
  * VexCL integration
  * More logical operators
  * Many more unit tests
  * Many assorted bug fixes

0.0.0
  * DualNumber recursive automatic differentiation
  * NumberVector, NumberArray operator overloading
  * Recursive gradients, tensors, etc.