forked from roystgnr/MetaPhysicL
-
Notifications
You must be signed in to change notification settings - Fork 3
/
NEWS
122 lines (100 loc) · 4.13 KB
/
NEWS
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
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.