Version 8.7 (May 2001)

  1. The first version

Version 9.3 (January 2004)

  1. Each scripts (*.mr) for running MULTILEVELRATE jobs are modified. The script automatically make a scratch directory so that all intermediate files created by MULTILEVELRATE will be put into this scratch directory which is separate from the input files. At the end of the calculations, the MULTILEVELRATE output files are copied from the scratch directory to the directory containing the input files. The MULTILEVEL input and output files are moved to a directory called MRTEMPFILES for any possible debugging. Then the temporary scratch directory containing unnecessary intermediate files is deleted. It makes the end user much easier to manage the input and output files by running MULTILEVELRATE jobs through this modified script.

  3. Input sections *MRREACT1, *MRREACT2, *MRPROD1, *MRPROD2, *MRSTART, *MRWELLR, and *MRWELLP are added to the fu70 file in order to specify charges for each corresponding species by the keyword CHARGE. For a MULTILEVELRATE calculation in which the stationary points are optimized through hooks, above information is ignored, regardless whether the species is neutral or not. When one defines the initial geometry as an optimized structure in POLYRATE (STATUS=2) for a charged species, this additional input of charges in fu70 is required. This corrects a bug for charged species in the previous version.

  5. A bug has been fixed in subroutine RSTATMR (hooks_mr.f), where a comma was missed in a print format statement. For some compilers, this bug could generate an error.

  7. The input keywords "SGLV" and "MTLV" (in section *MRGENERAL of fu70) are removed. By utilizing the summary file (ml.sum) printed out by MULTILEVEL, the special treatment required for a single level calculation does not exist any longer.

  9. Test runs for H2SOH, Al2, and CH3+ are added. The latter two cases are to show how to compute multilevel frequencies by using MULTILEVELRATE. The original CH5 test runs have been reorganized.