1 |
|
---|
2 | MPQC REVISION HISTORY
|
---|
3 |
|
---|
4 | Following is a brief summary of changes made in each release of MPQC.
|
---|
5 |
|
---|
6 | * 2006-03-22: MPQC-2.3.1
|
---|
7 | o MCSearch added for cubic interpolation during quasi-Newton line
|
---|
8 | searches.
|
---|
9 | o Added KMLYP method.
|
---|
10 | o Updated libtool to version 1.5.22.
|
---|
11 | o SumDenFunctional returns correct HF exchange coefficient when using
|
---|
12 | nested ACM functionals. This could change the results from certain,
|
---|
13 | uncommon inputs.
|
---|
14 | o Other minor bug fixes, enhancements, and documentation improvements.
|
---|
15 |
|
---|
16 | * 2005-10-16: MPQC-2.3.0
|
---|
17 | o Common Component Architecture interfaces for energies and
|
---|
18 | integrals are now available.
|
---|
19 | o Updated to a more recent version of the EMSL basis sets. This
|
---|
20 | will change results in many cases.
|
---|
21 | o Polarization consistent basis sets added.
|
---|
22 | o Test suite inputs and results have been moved from a separate
|
---|
23 | file into the main release.
|
---|
24 | o "make" targets are now available to automatically check results.
|
---|
25 | o "mpqcrun" can be used to simply starting MPQC runs, particulary
|
---|
26 | parallel runs.
|
---|
27 | o Complementary Auxiliary Basis Set (CABS) version of the MP2-R12
|
---|
28 | method [see Chem. Phys. Lett. 395, 190 (2004)].
|
---|
29 | o MP2-R12/A' method with split virtual space (virtual orbitals
|
---|
30 | constructed from a separate Gaussian basis).
|
---|
31 | o Novel Brillouin condition-free versions of the MP2-R12/A' method.
|
---|
32 | o A new exception infrastructure has been added.
|
---|
33 | o Numerous numerical improvements, compiler bug work-arounds,
|
---|
34 | and bug fixes.
|
---|
35 | o Checkpoint files made with earlier versions of MPQC utilizing
|
---|
36 | more than one processor will not work 2.3.0 or later versions.
|
---|
37 |
|
---|
38 | * 2004-12-18: MPQC-2.2.3
|
---|
39 | o Documention errors and omissions corrected.
|
---|
40 | o Command man pages generated automatically.
|
---|
41 | o bison and flex are no longer required.
|
---|
42 | o Will now compile with MPICH2.
|
---|
43 | o Maintain symmetry in optimizations of large molecules.
|
---|
44 | o Improved ease of builds within the source directory.
|
---|
45 | o Fixed density value computations.
|
---|
46 | o Port to Cray X1 (correctness of results not checked).
|
---|
47 | o Port to G5/OS X (correctness of results not checked).
|
---|
48 |
|
---|
49 | * 2004-04-28: MPQC-2.2.2
|
---|
50 | o Can now compile with GCC 3.4.0.
|
---|
51 | o Switched from ieee_{set,get}_fp_control to fenv functions.
|
---|
52 | o More architecture types recognized, including pentium4.
|
---|
53 | o The getnwbas.pl and parsenwbas.pl work with recent versions
|
---|
54 | of the EMSL Basis Set Database.
|
---|
55 | o Libtool flags are now accessible with sc-config.
|
---|
56 | o Psi 3 compatibility classes have been updated.
|
---|
57 |
|
---|
58 | * 2003-12-26: MPQC-2.2.1
|
---|
59 | o Fixed a problem where FORTRAN files caused builds of shared
|
---|
60 | libraries to fail.
|
---|
61 | o Duplicate identifiers for AMD64 shared library builds have been fixed.
|
---|
62 | o The SONAME of this release is 5:1:0. GCC 3.2 is the official C++ ABI.
|
---|
63 |
|
---|
64 | * 2003-12-04: MPQC-2.2.0
|
---|
65 | o (ABS) MP2-R12 method added (Edward Valeev).
|
---|
66 | o ARMCIMemoryGrp added.
|
---|
67 | o Better threading parallelism.
|
---|
68 | o The SONAME of this release is 5:0:0. GCC 3.2 is the official C++ ABI.
|
---|
69 |
|
---|
70 | * 2003-09-26: MPQC-2.1.5
|
---|
71 | o Fixes for shared library support (Michael Banck)
|
---|
72 | o Minor documentation updates.
|
---|
73 |
|
---|
74 | * 2003-07-22: MPQC-2.1.4
|
---|
75 | o Fixes a problem that occurs when optimizing with GCC 3.3.
|
---|
76 | o Now works with more recent versions of bison, flex, and autoconf.
|
---|
77 | o Upgraded libtool to 1.5 (this requires recent versions of
|
---|
78 | automake/autoconf).
|
---|
79 | o Bug in the charge computation for Extended Huckel is fixed.
|
---|
80 | o sc-config --cppflags should always give the correct include path now.
|
---|
81 | o The SONAME of this release is 4:1:0. GCC 3.2 is the official C++ ABI.
|
---|
82 |
|
---|
83 | * 2003-01-10: MPQC-2.1.3
|
---|
84 | o Fixes a problem that occurred with recent versions of bison. Note
|
---|
85 | however that a new problem has emerged in bison 1.875 which is not
|
---|
86 | fixed in this release.
|
---|
87 | o Applied changes to make RPM's easier to build.
|
---|
88 | + The installroot=path assignment can be given on the make command
|
---|
89 | line to specify a temporary install root.
|
---|
90 | + --with-sc-includedir=dir can be used to specify an installation
|
---|
91 | directory for the include files.
|
---|
92 |
|
---|
93 | * 2002-10-14: MPQC-2.1.2
|
---|
94 | o Bumped SONAME to 4:0:0. GCC 3.2 is the official C++ ABI for
|
---|
95 | 4:0:0. See the --with-build-id configuration documentation for more
|
---|
96 | info.
|
---|
97 | o Remove use of MPI-2 1-sided as this has proved inadequate.
|
---|
98 | o Minor documentation updates.
|
---|
99 |
|
---|
100 | * 2002-04-21: MPQC-2.1.1
|
---|
101 | o Support recent versions of bison and flex.
|
---|
102 |
|
---|
103 | * 2002-04-18: MPQC-2.1.0
|
---|
104 | o Everything is now in the "sc" namespace.
|
---|
105 | o Fixed a problem causing redundant output lines with gcc3.
|
---|
106 | o Added extended Huckel theory guesses.
|
---|
107 | o Added configure options:
|
---|
108 | + --enable-always-use-mpi
|
---|
109 | + --with-default-parallel={mpi2,mtmpi}
|
---|
110 | o Begin conversion to use exceptions instead of abort.
|
---|
111 | o Improved MP2 gradient code performance.
|
---|
112 | o Greater than 2GB support for memory keywords.
|
---|
113 | o Symbolic notation for memory keyword supported. Examples: 2KB, 1MB,
|
---|
114 | 1.3GB, 8KIB, 1MIB, 6GIB.
|
---|
115 | o Thread safety issues addressed in MessageGrp classes.
|
---|
116 | o Obsolete code removed.
|
---|
117 | o Fixed accuracy problem affecting mainly single point DFT gradients.
|
---|
118 |
|
---|
119 | * 2002-03-01: MPQC-2.0.4
|
---|
120 | o Fixed a checkpoint/restart bug introduced in 2.0.2.
|
---|
121 |
|
---|
122 | * 2002-02-17: MPQC-2.0.3
|
---|
123 | o Problems using newer versions of bison were fixed. Versions up to
|
---|
124 | 1.33 should work.
|
---|
125 | o Avoid duplicate MemoryGrp constructor calls.
|
---|
126 |
|
---|
127 | * 2002-01-24: MPQC-2.0.2
|
---|
128 | o Several configure problems were fixed.
|
---|
129 | o The keep_guess_wavefunction option was added. This is used to allow
|
---|
130 | the guess wavefunction to be reused for lower symmetry displacements
|
---|
131 | during a frequency calculation.
|
---|
132 | o Unrestricted wavefunctions will now correctly recompute occupations
|
---|
133 | when displaced into lower symmetry.
|
---|
134 | o A stack overflow for very large calculations was fixed.
|
---|
135 |
|
---|
136 | * 2002-01-08: MPQC-2.0.1
|
---|
137 | o Incorporate some of Michael Banck's Debian porting work.
|
---|
138 | o Improve usability on parallel machines without pthreads.
|
---|
139 | o Minor portability enhancements.
|
---|
140 |
|
---|
141 | * 2001-10-04: MPQC-2.0.0
|
---|
142 | o Add support for KAI KCC.
|
---|
143 | o Put template instantiations in libraries for Compaq C++.
|
---|
144 |
|
---|
145 | * 2001-08-31: MPQC-2.0.beta.6 html build 2
|
---|
146 | o Inheritance and collaboration diagrams added to documentation.
|
---|
147 |
|
---|
148 | * 2001-08-29: MPQC-2.0.beta.6
|
---|
149 | o Make sure MPI_Finalize gets called.
|
---|
150 | o Fixed scpr.
|
---|
151 | o Fixed parallel DFT printing.
|
---|
152 | o Added --enable-production configure option.
|
---|
153 | o Improved MP2 performance.
|
---|
154 |
|
---|
155 | * 2001-07-10: MPQC-2.0.beta.5
|
---|
156 | o Added Gram-Schmidt orthogonalization of basis functions.
|
---|
157 | o Changed the following basis sets to use 5D instead of 6D:
|
---|
158 | 6-311++G(2d,2p), 6-311++G(3df,3pd), 6-311++G**.
|
---|
159 | o Fixed bugs in MTMPIMemoryGrp exposed by MPI/Pro 1.6.3.
|
---|
160 | o Improved multi-threaded MP2 code.
|
---|
161 |
|
---|
162 | * 2001-04-06: MPQC-2.0.beta.4
|
---|
163 | o Major API changes
|
---|
164 | + Use RTTI for dynamic casts
|
---|
165 | + Uses smart pointer templates instead of macros
|
---|
166 | + Requires ISO standard C++ compiler
|
---|
167 | + May require patched flex
|
---|
168 | o Clean up geometries when symmetry is specified
|
---|
169 |
|
---|
170 | * 2001-03-08: MPQC-2.0.beta.3
|
---|
171 | o Object directory can be a subdirectory of the source directory.
|
---|
172 | o Documentation updates.
|
---|
173 | o --enable-stl fixed for standard C++ compilers.
|
---|
174 |
|
---|
175 | * 2001-03-03: MPQC-2.0.beta.2
|
---|
176 | o New, simplified input format is supported in addition to the
|
---|
177 | objected-oriented input.
|
---|
178 | o Many improvements to the Density Functional Theory code. It is now
|
---|
179 | ready for production use.
|
---|
180 | o Documentation converted to doxygen.
|
---|
181 | o Supports ISO 14882 standard C++. MPQC now might not work with some
|
---|
182 | older compilers.
|
---|
183 | o Nonstandard symmetry frames now work.
|
---|
184 | o All HF and DFT objects can use multi-threading.
|
---|
185 |
|
---|
186 | * 2001-02-25: MPQC-1.2.5
|
---|
187 | o "make install_devel" fixed for unified source and object dirs.
|
---|
188 | o Symmetry detection in presence of ghost atoms fixed.
|
---|
189 | o GCC 2.95.2 now can be used to compile MPQC. GCC 3.0 and above will
|
---|
190 | not work (this will be fixed in next major release).
|
---|
191 | o Nonstandard symmetry frames do not work. They are detected and will
|
---|
192 | cause an abort (this will be fixed in next major release).
|
---|
193 | o "make -j2" problems fixed.
|
---|
194 |
|
---|
195 | * 2000-09-07: MPQC-1.2.4
|
---|
196 | o MTMPIMessageGrp will be linked into MPQC if threads and MPI are
|
---|
197 | available.
|
---|
198 |
|
---|
199 | * 2000-01-31: MPQC-1.2.3
|
---|
200 | o install_devel will now install scconfig.h
|
---|
201 |
|
---|
202 | * 1999-11-04: MPQC-1.2.2
|
---|
203 | o Turned off core dumps. These could cause problems on the IBM SP.
|
---|
204 |
|
---|
205 | * 1999-10-24: MPQC-1.2.1
|
---|
206 | o Fix minor problem preventing a smooth compile on AIX.
|
---|
207 |
|
---|
208 | * 1999-10-20: MPQC-1.2
|
---|
209 | o The 1.1 input files will not work with 1.2
|
---|
210 | o Ported to IRIX and IRIX64
|
---|
211 | o Documentation improvements
|
---|
212 | o "make interface" no longer needed
|
---|
213 |
|
---|
214 | * 1999-08-19: MPQC-1.2alpha6
|
---|
215 | o Use DOC++ for all documentation
|
---|
216 | o Bug fixes and ports to glibc 2.1 and gcc 2.95
|
---|
217 | o Canonical orthogonalization option
|
---|
218 |
|
---|
219 | * 1998-10-23: MPQC-1.2alpha5
|
---|
220 | o Multi-threaded CLHF and MP2 gradients
|
---|
221 | o Removed all use of libg++
|
---|
222 |
|
---|
223 | * 1998-05-25: MPQC-1.2alpha4
|
---|
224 | o Save/restore and result accuracy bug fixes
|
---|
225 | o Reducing memory requirement for integrals
|
---|
226 | o configure improvements
|
---|
227 |
|
---|
228 | * 1998-04-03: MPQC-1.2alpha3
|
---|
229 | o Faster integrals with arbitrary angular momentum
|
---|
230 | o Finite displacement Hessian can be used as guess Hessian
|
---|
231 | o Checkpoint override capability added
|
---|
232 |
|
---|
233 | * 1997-10-30: MPQC-1.2alpha2
|
---|
234 | o Fix a bug affecting optimizations
|
---|
235 |
|
---|
236 | * 1997-10-29: MPQC-1.2alpha1
|
---|
237 | o Checkpoint files have directory information that can be used to
|
---|
238 | selectively restore objects
|
---|
239 | o Class hierarchy changed under SCF
|
---|
240 | o The molecule class was cleaned up
|
---|
241 | o Many other bug fixes and enhancements
|
---|
242 |
|
---|
243 | * 1997-6-11: MPQC-1.1.2
|
---|
244 | o Fixed SYSV IPC configuration problem on sgi-irix platforms.
|
---|
245 | o Fixed sample inputs in HTML user manual.
|
---|
246 |
|
---|
247 | * 1997-5-23: MPQC-1.1.1
|
---|
248 | o GNU-CC specific construct removed.
|
---|
249 |
|
---|
250 | * 1997-5-22: MPQC-1.1.0
|
---|