4
4
# SPDX-License-Identifier: (Apache-2.0 OR MIT)
5
5
6
6
#=============================================================================
7
- # This file chronicles user-level changes to hypre, beginning with the most
8
- # recent release.
9
- #=============================================================================
7
+ # This file chronicles user-level changes, beginning with the most recent.
8
+ # =============================================================================
10
9
11
10
Version 2.16.0 released 2019/03/20
12
11
@@ -24,12 +23,16 @@ Version 2.16.0 released 2019/03/20
24
23
25
24
- Various bug fixes.
26
25
26
+ #====================================
27
+
27
28
Version 2.15.1 released 2018/10/19
28
29
29
30
- Added hypre version functions: HYPRE_Version(), HYPRE_VersionNumber()
30
31
31
32
- Various bug fixes.
32
33
34
+ #====================================
35
+
33
36
Version 2.15.0 released 2018/09/21
34
37
35
38
- Added new system interface routines to reduce copies and improve threading:
@@ -46,6 +49,8 @@ Version 2.15.0 released 2018/09/21
46
49
47
50
- Various bug fixes.
48
51
52
+ #====================================
53
+
49
54
Version 2.14.0 released 2018/03/16
50
55
51
56
- New GPU release uses CUDA or OpenMP4.5. AMG setup phase is not yet ported.
@@ -56,6 +61,8 @@ Version 2.14.0 released 2018/03/16
56
61
57
62
- Various bug fixes.
58
63
64
+ #====================================
65
+
59
66
Version 2.13.0 released 2017/10/20
60
67
61
68
- Added new multigrid reduction (MGR) solver
@@ -74,6 +81,8 @@ Version 2.13.0 released 2017/10/20
74
81
75
82
- Fixed Windows Visual Studio compile (version 2017)
76
83
84
+ #====================================
85
+
77
86
Version 2.12.1 released 2017/09/29
78
87
79
88
- Added support for single and quad precision floating point numbers
@@ -82,12 +91,16 @@ Version 2.12.1 released 2017/09/29
82
91
83
92
- Various bug fixes
84
93
94
+ #====================================
95
+
85
96
Version 2.12.0 released 2017/05/02
86
97
87
98
- Added GPU support to hypre. The Struct and SStruct code can use CUDA, RAJA,
88
99
or KOKKOS. The ParCSR code uses CUDA. The BoomerAMG setup phase is not yet
89
100
ported to the GPU. This release uses unified memory.
90
101
102
+ #====================================
103
+
91
104
Version 2.11.2 released 2017/03/13
92
105
93
106
- Changed the defaults in hypre to HMIS with ext+i(4) interpolation
@@ -98,6 +111,8 @@ Version 2.11.2 released 2017/03/13
98
111
99
112
- Various bug fixes
100
113
114
+ #====================================
115
+
101
116
Version 2.11.1 released 2016/06/09
102
117
103
118
- Fixed one more bug related to SStructSetSharedPart and SetNeighborPart
@@ -115,6 +130,8 @@ Version 2.11.1 released 2016/06/09
115
130
- Updated the AME and LOBPCG interfaces to work correctly with both absolute and
116
131
relative tolerances.
117
132
133
+ #====================================
134
+
118
135
Version 2.11.0 released 2016/03/28
119
136
120
137
- Added a new function hypre_BoomerAMGSetKeepTranspose, which will store the
@@ -147,6 +164,8 @@ Version 2.11.0 released 2016/03/28
147
164
148
165
- Removed Babel from the release
149
166
167
+ #====================================
168
+
150
169
Version 2.10.1 released 2015/09/11
151
170
152
171
- Fixed various bugs in AMG and the IJ interface.
@@ -160,6 +179,8 @@ Version 2.10.1 released 2015/09/11
160
179
for the nonGalerkin version, with HYPRE_BoomerAMGSetNonGalerkinTol and
161
180
HYPRE_BommerAMGSetLevelNonGalerkinTol
162
181
182
+ #====================================
183
+
163
184
Version 2.10.0b released 2015/01/22
164
185
165
186
- Added new interpolation that allows the incorporation of near null-space
0 commit comments