Amber的CPU并行(intel优化)和GPU加速安装

帮购买的朋友安装,系统centos6.8,显卡750ti,cpu:intel

#cpu
cd /opt/biosoft/amber18
export MKL_HOME=/opt/intel 
export PATH=/opt/intel/bin/:$PATH
source /opt/intel/impi/2017.4.259/bin64/mpivars.sh
export PATH=/opt/intel/impi/2017.4.259/intel64/bin/:$PATH
export PATH=/opt/intel/:$PATH
sudo cp compilers_and_libraries_2017.7.259/linux/mkl/include/fftw include/ -R 
./configure -intelmpi intel
test -f /opt/biosoft/amber18/amber.sh && source /opt/biosoft/amber18/amber.sh
#sudo ln -s /opt/intel/compilers_and_libraries_2017.7.259/linux/mkl/lib/intel64_lin/ /opt/intel/lib/intel64/ 
sudo ln -s /opt/intel/compilers_and_libraries_2017.7.259/linux/mkl/lib/intel64_lin/* /opt/intel/lib/intel64/ 
make install
source /opt/intel/compilers_and_libraries_2017.7.259/linux/bin/ifortvars.sh intel64
make clean
./configure -openmp -intelmpi intel #amber16之后,由于amber18不建议用18版本的intel编译器,建议是用17或者16的编译器,所以需要安装老版本的intel编译器
test -f /opt/biosoft/amber18/amber.sh && source /opt/biosoft/amber18/amber.sh
make install
#gpu 事先安装好nvidia驱动和cuda
export CUDA_HOME=/usr/local/cuda-9.2
export LD_LIBRARY_PATH="/usr/local/cuda-9.2/lib:${LD_LIBRARY_PATH}"
source ~/.bashrc
#boost安装为了解决如下的库链接不合适问题
#Testing the g++ compiler with boost:
#     g++ -fPIC  testp.cpp -lboost_thread -lboost_system -o testp
#./testp: error while loading shared libraries: libboost_thread.so.1.57.0: canno
#
#Warning: Could not link to boost. Ensure boost-thread, boost-system and boost h
#Warning: Failed command:
#     g++ -fPIC  -lboost_thread -lboost_system -o testp testp.cpp
#Warning: Check boost.compile.err for error details.
#Warning: Memembed functionality in packmol_memgen not available.

wget http://sourceforge.net/projects/boost/files/boost/1.57.0/boost_1_57_0.tar.gz
tar -zxf boost_1_57_0.tar.gz
./bootstrap.sh --prefix=/opt/sysoft/boost-1.57
./b2 install

tar -xf AmberTools18.tar.bz2 -C /opt/biosoft/
tar -xf Amber18.tar.bz2 -C /opt/biosoft/
cd /opt/biosoft/amber18/
#./configure -cuda gnu
#./configure -cuda --prefix /opt/biosoft/amber18/amber_software_bin/ gnu
./configure -cuda gnu
test -f /opt/biosoft/amber18/amber.sh && source /opt/biosoft/amber18/amber.sh
make install
#由于nvcc调用c++编译时不支持老版本gcc-4.4.7,于是使用高版本gcc-4.9.3,需要先安装一个高版本的gcc
#configure报错libquadmath.so.0: cannot open
export LD_LIBRARY_PATH=/opt/sysoft/gcc-4.9.3/lib/:$LD_LIBRARY_PATH  
export LD_LIBRARY_PATH=/opt/sysoft/gcc-4.9.3/lib64/:$LD_LIBRARY_PATH 
export C_INCLUDE_PATH=/opt/sysoft/gcc-4.9.3/include/:$C_INCLUDE_PATH 
export CPLUS_INCLUDE_PATH=/opt/sysoft/gcc-4.9.3/include/:$CPLUS_INCLUDE_PATH 

make test
#error while loading shared libraries: libcurand.so.9.2: cannot open shared object file: No such file or directory
ldd pmemd.cuda
export LD_LIBRARY_PATH=/usr/local/cuda-9.2/lib64${LD_LIBRARY_PATH:+:${LD_LIBRARY_PATH}}
export LD_LIBRARY_PATH="/usr/local/cuda-9.2/lib64:${LD_LIBRARY_PATH}"

nvidia-smi

软件购买:http://ambermd.org/

此条目发表在Molecular Dynamics分类目录。将固定链接加入收藏夹。

发表评论

电子邮件地址不会被公开。 必填项已用*标注

To create code blocks or other preformatted text, indent by four spaces:

    This will be displayed in a monospaced font. The first four 
    spaces will be stripped off, but all other whitespace
    will be preserved.
    
    Markdown is turned off in code blocks:
     [This is not a link](http://example.com)

To create not a block, but an inline code span, use backticks:

Here is some inline `code`.

For more help see http://daringfireball.net/projects/markdown/syntax

Protected with IP Blacklist CloudIP Blacklist Cloud