New Features
In addition to several important bug fixes, version 4.2 of the
Hibridon code has the following new features:
- An updated module (hitensor.f) for the determination of tensor cross sections
and M-dependent cross sections.
- A component (hibound.f) for the determination of weakly-bound complexes involving
atomic and/or molecular radicals, provided that the interaction potential can be
expanded in the same form as used for molecular scattering.
- An option, when linking the code, to exclude certain functions in exchange for an
ability to handle a larger number of channels
- Documentation of the option allowing the restart of a long-running calculation after
a power failure or system crash
The on-line (doc/hibrid.hlp) and .html-based (doc/index.html) help facilities have been
revised to include help on these items as well as examples.
Version 4.2 of the code includes:
- Extension to Apple's G4 and G5 processor workstations running OSX 10.3 (Panther). In order to compile
and link on Apple workstations, you will need to have:
- The IBM XL Fortran Compiler (XLF) available from Absoft
- The Mac OS X Developer Tools package, which includes gcc Version 3.3.
The package is available from the Apple Developer Connection
Website. In particular, Xcode Tools v1.0 or later is required.
- Replacement of many matrix-matrix: routines with calls to version 3.0
Lapack routines.
In addition to the changes mentioned above for version 4.2, version 4.3 of the Hibridon code has
the following new features:
- The possibility of doing calculations at up to 25 different collision energies, which
may be chosen on a regularly spaced linear grid.
- Two possibilities for formatting the integral cross section files arising from the commands
PRINTC and INTCRS.
- A new command, SYSCONF (and an equivalent shell command hib_sysconfig) which prints out the key details
of your current operating system.
The on-line (doc/hibrid.hlp) and .html-based (doc/index.html) help facilities have been
revised to include help on these items as well as examples.
Version 4.3 of the code has been developed for (and tested on) the following platforms:
- Apple's G4 and G5 processor workstations running OSX 10.3 (Panther) and 10.4 (Tiger). In order to compile
and link on Apple G4 and G5 workstations, you will need to have:
- The IBM XL Fortran Compiler (XLF)
available from Absoft Note, despite the indication
on Absoft's web page that IBM XL Fortran v8.1 for Mac OS X is not supported on Mac OS X Version 10.4 - Tiger,
the Hibridon package does install, compile, and execute correctly under OSX 10.4.10 and 10.4.11 with XLF8.1. Hibridon is no longer actively
maintained on Apple PowerPC platforms.
- The current version of Mac OS X Developer Tools package.
Under OS X 10.5 and 10.6 this package is available from the Apple Developer Connection
Website. Under System 10.7 (Lion) the Xcode Tools package is available directly from the App Store
.
- Apple Intel platforms running OSX 10.4 (Tiger) and 10.5.1 (Leopard). In order to compile
and link on Apple Intel platforms, you will need to have:
- The
Intel Fortran Compiler (ifort).
- The Mac OS X Developer Tools package.
The package is available from the Apple Developer Connection
Website. In particular, the current version of Xcode Tools is required.
Under System 10.7 (Lion) this is available directly from the App Store
- Intel Xeon (x86) based platforms running under Linux (installation under Red Hat v. 4 has
been explicitly tested).
Compilation, linking, and execution of Hibridon v. 4.3 has been successful with
- The Intel ifort compiler and MKL mathematics library
- The Portland Group pgf95 compiler and ACML mathematics library
The major computational effort in calculations with the Hibridon code is associated with
a series of matrix-matrix (BLAS3) operations (matrix inversions, diagonizations, and multiplications), which
are carried out by calls to the LAPACK routines DGERF, DGETRI, DSYEVR, and DGEMM.
The latest version
of Intel's (MKL math library as well as
of Apple's veclib (/System/Library/Frameworks/Accelerate.framework/Versions/Current) include, automatically,
parallelized versions of these LAPACK routines, so that considerable performance enhancement can be
achieved when running on multi-cpu, multi-core workstations. A more in-depth discussion is given
on a separate page.
Go to
Installation | Benchmarks |
Downloading | Memory Limits |
Restarting | Timing | Parallelism | Linking | Hibridon Help