• バックアップ一覧
• 差分 を表示
• 現在との差分 を表示
• ソース を表示
• QE_tutorial/install_smith_new_6.6 へ行く。
  • 1 (2020-10-14 (水) 18:55:09)
  • 2 (2020-10-14 (水) 19:30:27)
  • 3 (2020-10-23 (金) 01:01:52)
  • 4 (2020-10-27 (火) 14:38:50)
  • 5 (2020-10-30 (金) 00:25:29)
  • 6 (2020-10-30 (金) 17:40:40)
  • 7 (2020-10-31 (土) 23:15:49)
  • 8 (2020-11-06 (金) 19:11:42)
  • 9 (2020-11-10 (火) 15:12:54)
  • 10 (2020-11-10 (火) 23:50:57)
  • 11 (2021-04-10 (土) 16:24:32)

---

Installation †

Here how to compile Quantum-ESPRESSO 6.6 on smith@Morikawa Group, Osaka University.

First of all, make sure that the modules are loaded properly. In this tutorial, the following modules are used:

intel/2020.2.254
intelmpi/2020.2.254
python/3.8

To load the above modules, add the following in ${HOME}/.bashrc:

module load intel/2020.2.254
module load intelmpi/2020.2.254
module load python/3.8

Make sure that old settings for intel compiler are deleted/commented out.

Let us first make directories for source codes, pseudopotentials, and calculations for Quantum-ESPRESSO in the home directory, for e.g.

$ mkdir -p QE/src
$ mkdir -p QE/pseudo
$ mkdir -p QE/out

Obtaining the code †

In the ${HOME}/QE/src directory, type

$ git clone -b 'qe-6.6' --depth 1 https://gitlab.com/QEF/q-e.git qe-6.6

If "qe-6.6" is not specified, "q-e" is generated. One can also do

$ wget https://gitlab.com/QEF/q-e/-/archive/qe-6.6/q-e-qe-6.6.tar.bz2

and untar

$ tar jxf q-e-qe-6.6.tar.bz2
$ mv q-e-qe-6.6 qe-6.6

Configure and build †

QE †

In qe-6.6, execute the following

$ MPIF90=mpiifort DFLAGS='-D__OPENMP -D__FFTW3 -D__MPI -D__SCALAPACK' SCALAPACK_LIBS='-L${MKLROOT}/lib/intel64 -lmkl_scalapack_lp64 -lmkl_blacs_intelmpi_lp64' ./configure --enable-openmp

The following option may be added to FFLAGS* in "make.inc"

-xSSE4.2 -axCOMMON-AVX512,CORE-AVX512,CORE-AVX2,CORE-AVX-I,AVX

Then type Furthermore, the following should be added (not to be overwritten) to IFLAGS in "make.inc"

-I$(MKLROOT)/include/fftw

If you want to use BEEF(-vdW) functional, add "-Duse_beef" to the CPP option in "make.inc as

MANUAL_DFLAGS  = -Duse_beef

Then type

$ make pw

Other programs can be found by just typing

$ make

I usually build PW as well as PP by typing

$ make pw pp

If you like to perform phonon calculations do

$ make ph

Yambo †

According to the makefile

$ make yambo

to build yambo, but it failed (see an alternative way (experimental)). Alternatively, I build yambo as follows:
First download the latest stable version of yambo:

$ wget https://github.com/yambo-code/yambo/archive/4.5.2.tar.gz

and untar

$ tar zxvf 4.5.2.tar.gz

Go to the yambo directory

$ cd yambo-4.5.2

and configure: (without scalapack)

PFC="mpiifort" \
FC="ifort" \
F77="ifort" \
./configure \
--enable-mpi \
--enable-open-mp \
--enable-dp \
--with-lapack-libs="-mkl=parallel" \
--with-blas-libs="-mkl=parallel" \
--with-fft-libs="-mkl=parallel" \
--with-iotk-path="/home/ikutaro/QE/src/qe-6.6/iotk" \

(with scalapack)

$ PFC="mpiifort" \
FC="ifort" \
F77="ifort" \
./configure \
--enable-mpi \
--enable-open-mp \
--enable-dp \
--with-iotk-path="${HOME}/QE/src/qe-6.6/iotk" \
--with-blas-libs="-lmkl_intel_lp64 -lmkl_intel_thread -lmkl_core -liomp5 -lpthread -lm" \
--with-lapack-libs="-lmkl_intel_lp64 -lmkl_intel_thread -lmkl_core -liomp5 -lpthread -lm" \
--with-fft-includedir="-I${MKLROOT}/include/fftw" \
--with-fft-libs="-lmkl_intel_lp64 -lmkl_intel_thread -lmkl_core -liomp5 -lpthread -lm" \
--with-blacs-libs="-lmkl_scalapack_lp64 -lmkl_blacs_intelmpi_lp64 -lmkl_core -liomp5 -lpthread -lm" \
--with-scalapack-libs="-lmkl_scalapack_lp64 -lmkl_blacs_intelmpi_lp64 -lmkl_core -liomp5 -lpthread -lm"

NOTE that even though libxc is specified explicitly, the internal ones are built (I hope to figure out why in near future).
Finally we are able to build yambo and tools as:

$ make yambo ypp p2y

and the executables can be found in the bin/ directory.