type
status
date
slug
tags
summary
category
password
icon
本所有编译都是在/OPT下面编译，共享目录
3.下载lammps 选择第一个稳定版下载 https://lammps.sandia.gov/download.html
4.编译fftw
cd /opt
mkdir fftw3.3.8
wget http://www.fftw.org/fftw-3.3.8.tar.gz
tar -xvf fftw-3.3.8.tar.gz
cd fftw-3.3.8/
./configure --prefix=/opt/fftw3.3.8 --enable-shared=yes
make -j 4 && make install
5.编译mpich
cd /opt
mkdir mpich3.3.2
wget http://www.mpich.org/static/downloads/3.3.2/mpich-3.3.2.tar.gz
cd mpich-3.3.2
./configure --prefix=/opt/mpich3.3.2
make -j 4 && make install
6.编译lammps
将下载好的lammps解压，进入文件内
cd /opt/lammps-3Mar20/src/MAKE
vim Makefile.mpi
将下面一行注释，如图
LMP_INC = -DLAMMPS_GZIP -DLAMMPS_MEMALIGN=64 # -DLAMMPS_CXX98
修改保存退出
cd ../src
加入模块
make package-status //查看lammps可用模块
make yes-all
make no-lib
make no-ext
make -j 4 mpi
环境变量
export PATH=/opt/mpich3.3.2/bin:/opt/fftw3.3.8/bin:/opt/lammps-3Mar20/src:$PATH
export LD_LIBRARY_PATH=/opt/fftw3.3.8/lib:/opt/mpich3.3.2/lib:$LD_LIBRARY_PATH
以上就全部编译完成
------分割--------
然后进入算例mpirun -np 8 lmp_mpi < in.friction
算例正常运行
如果需要安装gpu版
cd /opt/lammps-3Mar20/lib/gpu
vim Makefile.linux
CUDA_HOME=/usr/local/cuda-11.1
下面参数修改请看**CUDA**维基百科
Turing hardware
CUDA_ARCH = -arch=sm_75
修改完成保存退出
make -f Makefile.linux
cd /opt/lammps-3Mar20/src
make yes-gpu
make -j mpi
计算12核加gpu
mpirun -np 12 lmp_mpi -sf gpu -pk gpu 1 -in in.friction
关于编译lammps+plumed
需下载lammps-stable_7Aug2019的版本
cd src
make lib-plumed args="-b"
make yes-all
make no-lib
make no-ext
make yes-plumed或make yes-user-plumed
make -j 8 mpi
- Author:NotionNext
- URL:https://tangly1024.com/article/ab4281ac-f0a7-47de-99cf-608a7dfbde24
- Copyright:All articles in this blog, except for special statements, adopt BY-NC-SA agreement. Please indicate the source!
