orca410频率、优化计算出错termination in GTOInt

快别笑了啦

超算上进行orca410(openmpi-3.1.3)的频率、优化计算时出错,百思不得其解，请教大家！：
体系19个原子（CHO）freq.inp
#
%pal nprocs 32 end
!UKS wb97x-d3 cc-pvdz freq
* xyz 0 2

*

log文件出现:
ORCA finished by error termination in GTOInt
Calling Command: mpirun -np 32  /public/home/dhua/soft/orca_4_1_0/orca_gtoint_mpi freq.int.tmp freq
[file orca_tools/qcmsg.cpp, line 453]:
  .... aborting the run


.err文件
[b11r2n15:07499] *** Process received signal ***
[b11r2n15:07499] Signal: Segmentation fault (11)
[b11r2n15:07499] Signal code: Address not mapped (1)
[b11r2n15:07499] Failing at address: (nil)
[b11r2n15:07499] [ 0] /lib64/libpthread.so.0(+0xf5d0)[0x2b7e050f25d0]
[b11r2n15:07499] *** End of error message ***
[file orca_tools/qcmsg.cpp, line 453]:
  .... aborting the run


real        0m0.917s
user        0m0.027s
sys        0m0.128s


提交脚本
#!/bin/bash
#SBATCH -J yhj
#SBATCH -p normal
#SBATCH -N 1
#SBATCH -n 32
#SBATCH --error=%J.err
#SBATCH --output=%J.out

module load compiler/intel/2020.1.217

export PATH=/public/home/dhua/soft/openmpi-3.1.3/bin:/public/home/dhua/soft/orca_4_1_0:$PATH
export LD_LIBRARY_PATH=/public/home/dhua/soft/openmpi-3.1.3/lib:/public/home/dhua/soft/orca_4_1_0:$LD_LIBRARY_PATH
export orcapath=/public/home/dhua/soft/orca_4_1_0

for INPUT in *.inp
do
        echo "ORCA job start at" 'date'
        time $orcapath/orca $INPUT > log
done

zjxitcc

你这ORCA也太上古了，为何不换成ORCA 5.0.4试试。另外，最好指定一下每核内存%maxcore

快别笑了啦

zjxitcc 发表于 2023-12-25 21:21
你这ORCA也太上古了，为何不换成ORCA 5.0.4试试。另外，最好指定一下每核内存%maxcore

好的谢谢您，之前的体系都是用0RCA 4.1.0算的，就沿用了这个版本。
更换了超算有的mpi后成功运行啦