|
|
今天收到高斯公司邮寄的G09 E.01版本,看看和D01的不同之处,好像没有太大差别。和当初大家一起讨论的有所出入。其中,好像第4项,对激发态计算有所改进,好像进步不大,懂的给解释解释,这些改进能解决哪些问题?不翻译了,大家直接看吧。
*Feature and Usage notes: Rev E.01 This revision is primarily to fix a variety of bugs reported by users or which were uncovered by the example jobs for the 3rd edition ofExploring Chemistry with Electronic Structure Methods.
♦ Formatted checkpoint files now include the version of G09 used and also a job status flag, so that GUIs can determine whether ajob is still running, failed or completed successfully. The number of imaginary frequencies is also included for frequency jobs.
♦ EOMCC=ListWindow now reads the window only once instead of twice.
♦ Rigid (unrelaxed) scans with large energies now generate a table of relative energies rather than ****’s.
♦ The TDHF excited states method now defaults to using the MO basis, the same default as for CIS.
♦ A fix for the reference atoms used in Hirshfeld charges will change the reported charges slightly.
♦ The overhead in using Linda parallelism with ECPs was removed.
♦ The build procedures include compiling natively for Sandybridge (AVX) x86_64 processors, including an AVX-enabled ATLASlibrary. A Sandybridge/Haswell binary distribution is also available.
♦ The divide-and-conquer diagonalization and SVD routines work in a larger fraction of cases.
|
|