summaryrefslogtreecommitdiff
path: root/packages/hmatrix/CHANGELOG
diff options
context:
space:
mode:
authorAlberto Ruiz <aruiz@um.es>2014-05-08 08:48:12 +0200
committerAlberto Ruiz <aruiz@um.es>2014-05-08 08:48:12 +0200
commit1925c123d7d8184a1d2ddc0a413e0fd2776e1083 (patch)
treefad79f909d9c3be53d68e6ebd67202650536d387 /packages/hmatrix/CHANGELOG
parenteb3f702d065a4a967bb754977233e6eec408fd1f (diff)
empty hmatrix-base
Diffstat (limited to 'packages/hmatrix/CHANGELOG')
-rw-r--r--packages/hmatrix/CHANGELOG219
1 files changed, 219 insertions, 0 deletions
diff --git a/packages/hmatrix/CHANGELOG b/packages/hmatrix/CHANGELOG
new file mode 100644
index 0000000..3ddb4cb
--- /dev/null
+++ b/packages/hmatrix/CHANGELOG
@@ -0,0 +1,219 @@
10.16.0.0
2--------
3
4 * Added more organized reexport modules:
5 Numeric.HMatrix, Numeric.HMatrix.Data, Numeric.HMatrix.Devel
6 (The documentation is hidden for the other modules,
7 but they continue to be exposed and are not deprecated)
8
9 * added support for empty arrays
10
11 * join deprecated, use vjoin
12
13 * added (·) = cdot, which conjugates the first input vector
14 * added udot (unconjugated dot product)
15 * deprecated dot, use udot or (×) to keep the old behaviour
16
17 * added alternative overloaded multiplication operator (×) (or (<.>))
18 * (<.>) changed to infixr
19 * added Monoid instance for Matrix using matrix product
20
21 * improved build and konst
22
23 * improved linspace
24
25 * improved error messages
26 * more usage examples
27
28 * simplified LSDiv
29 * Plot functions moved to Numeric.LinearAlgebra.Util
30 * removed (!) (use (¦)), added (——)
31
320.15.2.0
33--------
34
35 * general pinvTol and improved pinv
36
370.15.1.0
38--------
39
40 * One-dimensional minimization
41
42 * Doubly-adaptive quadrature for difficult integrands
43
440.15.0.0
45--------
46
47 * Data.Packed.Foreign (additional FFI helpers)
48
49 * NFData instance of Matrix
50
51 * Unidimensional root finding
52
53 * In Numeric.LinearAlgebra.Util:
54 pairwise2D, rowOuters, null1, null1sym, size, unitary, mt, (¦), (?), (¿)
55
56 * diagBlock
57
58 * meanCov moved to Container
59
600.14.1.0
61--------
62
63 * In Numeric.LinearAlgebra.Util:
64 convolution: corr, conv, corr2, conv2, separable, corrMin
65 kronecker: vec, vech, dup, vtrans
66
670.14.0.0
68--------
69
70 * integration over infinite intervals
71
72 * msadams and msbdf methods for ode
73
74 * Numeric.LinearAlgebra.Util
75
76 * (<\>) extended to multiple right-hand sides
77
78 * orth
79
800.13.0.0
81--------
82
83 * tests moved to new package hmatrix-tests
84
850.11.2.0
86--------
87
88 * geigSH' (symmetric generalized eigensystem)
89
90 * mapVectorWithIndex
91
920.11.1.0
93--------
94
95 * exported Mul
96
97 * mapMatrixWithIndex{,M,M_}
98
990.11.0.0
100--------
101
102 * flag -fvector default = True
103
104 * invlndet (inverse and log of determinant)
105
106 * step, cond
107
108 * find
109
110 * assoc, accum
111
1120.10.0.0
113--------
114
115 * Module reorganization
116
117 * Support for Float and Complex Float elements (excluding LAPACK computations)
118
119 * Binary instances for Vector and Matrix
120
121 * optimiseMult
122
123 * mapVectorM, mapVectorWithIndexM, unzipVectorWith, and related functions.
124
125 * diagRect admits diagonal vectors of any length without producing an error,
126 and takes an additional argument for the off-diagonal elements.
127
128 * different signatures in some functions
129
1300.9.3.0
131--------
132
133 * flag -fvector to optionally use Data.Vector.Storable.Vector
134 without any conversion.
135
136 * Simpler module structure.
137
138 * toBlocks, toBlocksEvery
139
140 * cholSolve, mbCholSH
141
142 * GSL Nonlinear Least-Squares fitting using Levenberg-Marquardt.
143
144 * GSL special functions moved to separate package hmatrix-special.
145
146 * Added offset of Vector, allowing fast, noncopy subVector (slice).
147 Vector is now identical to Roman Leshchinskiy's Data.Vector.Storable.Vector,
148 so we can convert from/to them in O(1).
149
150 * Removed Data.Packed.Convert, see examples/vector.hs
151
1520.8.3.0
153--------
154
155 * odeSolve
156
157 * Matrix arithmetic automatically replicates matrix with single row/column
158
159 * latexFormat, dispcf
160
1610.8.2.0
162--------
163
164 * fromRows/fromColumns now automatically expand vectors of dim 1
165 to match the common dimension.
166 fromBlocks also replicates single row/column matrices.
167 Previously all dimensions had to be exactly the same.
168
169 * display utilities: dispf, disps, vecdisp
170
171 * scalar
172
173 * minimizeV, minimizeVD, using Vector instead of lists.
174
1750.8.1.0
176--------
177
178 * runBenchmarks
179
1800.8.0.0
181--------
182
183 * singularValues, fullSVD, thinSVD, compactSVD, leftSV, rightSV
184 and complete interface to [d|z]gesdd.
185 Algorithms based on the SVD of large matrices can now be
186 significantly faster.
187
188 * eigenvalues, eigenvaluesSH
189
190 * linearSolveLS, rq
191
1920.7.2.0
193--------
194
195 * ranksv
196
1970.7.1.0
198--------
199
200 * buildVector/buildMatrix
201
202 * removed NFData instances
203
2040.6.0.0
205--------
206
207 * added randomVector, gaussianSample, uniformSample, meanCov
208
209 * added rankSVD, nullspaceSVD
210
211 * rank, nullspacePrec, and economy svd defined in terms of ranksvd.
212
213 * economy svd now admits zero rank matrices and return a "degenerate
214 rank 1" decomposition with zero singular value.
215
216 * added NFData instances for Matrix and Vector.
217
218 * liftVector, liftVector2 replaced by mapVector, zipVector.
219