|
本帖最后由 cokie 于 2024-3-26 16:16 编辑
各位老师好,我在使用ORCA 5.0.3 win版本程序计算自旋-轨道耦合常数时,发生如下报错
- [file orca_tools/Tool-Linalg/qcrpasolv.cpp, line 785, Process 0]: Rebuilding of the RPA subspace failed because there are no old eigenvectors, aborting
- [file orca_tools/Tool-Linalg/qcrpasolv.cpp, line 785, Process 10]: Rebuilding of the RPA subspace failed because there are no old eigenvectors, aborting
- [file orca_tools/Tool-Linalg/qcrpasolv.cpp, line 785, Process 15]: Rebuilding of the RPA subspace failed because there are no old eigenvectors, aborting
- [file orca_tools/Tool-Linalg/qcrpasolv.cpp, line 785, Process 13]: Rebuilding of the RPA subspace failed because there are no old eigenvectors, aborting
- [file orca_tools/Tool-Linalg/qcrpasolv.cpp, line 785, Process 14]: Rebuilding of the RPA subspace failed because there are no old eigenvectors, aborting
- job aborted:
- [ranks] message
- [0-15] process exited without calling finalize
- ---- error analysis -----
- [0-15] on DESKTOP-NFN1R15
- C:\ORCA\orca_cis_mpi.exe ended prematurely and may have crashed. exit code 78
- ---- error analysis -----
- ORCA finished by error termination in CIS
- Calling Command: mpiexec -np 16 C:\ORCA\orca_cis_mpi.exe PXD_soc.cisinp.tmp PXD_soc
- [file orca_tools/qcmsg.cpp, line 465]:
- .... aborting the run
复制代码
我自己做过几个对比:更换泛函/基组(辅助基组)、增加nroots(之前看sob老师在某个帖子给过的可能有用的建议)、不使用TDA FALSE关键词、将并行核数改为12(每个核3500 MB,我的机器是16核64GB)均有上述报错。但我更换分子发现并没有上述报错,计算可以正常进行,所以应该也不是mpi.exe出了问题。但我不是很明白我的分子有什么特别之处。我把输入文件、输出文件和分子结构(用gaussian做过S1的优化)放下面了,希望能得到各位老师的帮助和解惑,非常感谢!
|
|