1 Installation

1.1 QE 6.7 + gcc 9.3 + OpenMPI 3.1.6 + OpenBlas 0.3.10 + FFTW3 (passed) + thermo_pw

tar -xvf qe-6.7-ReleasePack.tgz
cd qe-6.7
# download thermo_pw <https://github.com/dalcorso/thermo_pw/releases/download/1.4.1/thermo_pw.1.4.1.tar.gz>
tar -xvf thermo_pw.1.4.1.tar.gz
# Modify Modules/parameters.f90 and set npk=999999
# Modify PHonon/PH/phcom.f90 and set maxter = 500
# Modify PHonon/PH/lambda.f90 and set npk = 500

make veryclean
export FFTW_INCLUDE=/home/ywmu/backup/Src/comlib/fftw-3.3.8-gcc-4.8.5/include/ 
./configure --with-scalapack=yes FC=gfortran F90=gfortran F77=gfortran CC=gcc MPIF90=mpif90 BLAS_LIBS="/home/ywmu/backup/Src/comlib/gnu/libopenblas.a" SCALAPACK_LIBS="/home/ywmu/backup/Src/comlib/gnu/libscalapack.a"   FFT_LIBS="/home/ywmu/backup/Src/comlib/fftw-3.3.8-gcc-4.8.5/lib/libfftw3.a"  --prefix=/home/ywmu/backup/config/soft/qe.bin/qe68.gcc7.fftw
## Tips: 1. Add -ffpe-summary=none to FFLAGS
cd thermo_pw
make join_qe
cd .. 
nohup make all xspectra epw w90 want d3q  yambo thermo_pw &
tail -f nohup.out
make install

Mind: want will generate bands.x and dos.x, and they will cover ones generated by pp. So you should rename them.

1.2 QE 6.6 + gcc 9.3.0 + OpenMPI 3.1.6 + AMD AOCL (passed)

tar -xvf qe-6.6-ReleasePack.tgz
cd qe-6.6
# Modify Modules/parameters.f90 and set npk=999999
# Modify PHonon/PH/ phcom.f90 and set maxter = 500
export FFTW_INCLUDE=/opt/local/amd/2.2/amd-fftw/include 
./configure --with-scalapack=yes FC=gfortran F90=gfortran F77=gfortran CC=gcc MPIF90=mpif90 BLAS_LIBS="/opt/local/amd/2.2/amd-blis/lib/libblis.a" LAPACK_LIBS="/opt/local/amd/2.2/amd-libflame/lib/libflame.a" SCALAPACK_LIBS="/opt/local/amd/2.2/amd-scalapack/lib/libscalapack.a" FFT_LIBS="/opt/local/amd/2.2/amd-fftw/lib/libfftw3.a" --prefix=/home/ywmu/backup/Src/qe/bin/qe66.gcc9.amd
## Tips: 1. delete  BEEF_LIBS,BEEF_LIBS_SWITCH and -Duse_beef 2. Add -D__FFTW3 to DFLAGS 3. Add -ffpe-summary=none to FFLAGS 
nohup make pw ph pp pwcond neb cp ld1 tddfpt xspectra hp epw w90 yambo  d3q &
tail -f nohup.out
make install

1.3 pgi 18.10 (required) + OpenMPI 3.1.6 + MKL 2015 + Cuda 9.2 (GPU version passed)

Do not use -enable-openmp flag for configuration, errors happen easily!
Only pw.x ported on GPU!

tar -xvf qq-e-gpu-qe-gpu-6.5a2.tar.gz
cd q-e-gpu-qe-gpu-6.5a2
export PATH=/opt/pgi/linux86-64-llvm/18.10/bin:$PATH
export LD_LIBRARY_PATH=/opt/pgi/linux86-64-llvm/18.10/lib:$LD_LIBRARY_PATH
# load openmpi path 
# Computing power of 1080Ti: 60
export MKLROOT=/opt/intel15/composerxe/mkl
./configure --with-scalapack=no  FC=pgfortran F90=pgfortran F77=pgfortran CC=pgcc MPIF90=mpif90  BLAS_LIBS="-L${MKLROOT}/lib/intel64 -lmkl_intel_lp64 -lmkl_sequential -lmkl_core -lpthread -lm -ldl"  FFT_LIBS="/opt/local/fftw3.pgi18/lib/libfftw3.a" --prefix=/home/muyw/backup/qe.mkl  -with-cuda="/usr/local/cuda-9.2" --with-cuda-cc=60  --with-cuda-runtime=9.2  
make pw 
make install

1.3.1 Notes

Cuda 9.1/9.2/10.0 in Pgi compiler failed, so a standalone cuda-9.2 was used.
-with-cuda-cc=60 comes from the output of pgaccelinfo.
Only pw.x and phcg.x could be accelerated by GPU in 6.6a1 version.

1.4 intel MPI (high version needed)

tar -xvf q-e-qe-6.5.tar.gz
cd q-e-qe-6.5

. ~/.soft/envs/env_compiler oneapi
export FFTW_INCLUDE=${MKLROOT}/include/fftw
./configure --with-scalapack=intel F90=ifort F77=ifort CC=icc MPIF90=mpiifort  FFT_LIBS="${MKLROOT}/interfaces/fftw3xf/libfftw3xf_intel.a" --prefix=/home/ywmu/backup/Src/qe/bin.onepai     #-enable-openmp
nohup make pw ph pp pwcond neb cp  ld1 tddfpt xspectra hp epw w90 yambo  d3q &
make install

1.4.1 TYCloud

tar -xvf q-e-qe-6.8.tar.gz
cd q-e-qe-6.8

module remove mpi/hpcx/gcc-7.3.1
module load compiler/intel/2021.3.0
module load  mpi/intelmpi/2021.3.0
export FFTW_INCLUDE=${MKLROOT}/include/fftw
./configure --with-scalapack=intel F90=ifort F77=ifort CC=icc MPIF90=mpiifort  FFT_LIBS="/public/home/ywmu_sxu/.soft/libadd/staticlib/oneapi/libfftw3xf_intel.a" --prefix=/public/home/ywmu_sxu/source/bin68 
nohup make pw ph pp pwcond neb cp  ld1 tddfpt xspectra hp epw w90 yambo  d3q &
make install

1.5 QE 7.1 + ELPA 2021.11 + Intel onepai 2021.11.0 (test)

cd elpa-2021.11.001
FC=mpiifort  CC=icc ./configure --prefix=/home/ywmu/backup/Src/comlib/elpa FCFLAGS="-O3" CFLAGS="-O3" --enable-option-checking=fatal  SCALAPACK_LDFLAGS="-L$MKL_HOME/lib/intel64 -lmkl_scalapack_lp64 -lmkl_intel_lp64 -lmkl_sequential -lmkl_core -lmkl_blacs_intelmpi_lp64 -lpthread -lm -Wl,-rpath,$MKL_HOME/lib/intel64" SCALAPACK_FCFLAGS="-L$MKL_HOME/lib/intel64 -lmkl_scalapack_lp64 -lmkl_intel_lp64 -lmkl_sequential -lmkl_core -lmkl_blacs_intelmpi_lp64 -lpthread -I$MKL_HOME/include/intel64/lp64" --enable-avx2
make
make install 

cd q-e-qe-7.1
export FFTW_INCLUDE=${MKLROOT}/include/fftw
export ELPAROOT=/home/ywmu/backup/Src/comlib/elpa
export ELPAVER=elpa-2021.11.001
./configure F90=ifort F77=ifort CC=icc MPIF90=mpiifort  FFT_LIBS="${MKLROOT}/interfaces/fftw3xf/libfftw3xf_intel.a" --prefix=/home/ywmu/backup/Src/qe/bin.onepai --with-scalapack=intel -enable-parallel -enable-openmp --with-elpa-include="${ELPAROOT}/include/${ELPAVER}/modules" --with-elpa-lib="${ELPAROOT}/lib/libelpa.a" --with-elpa-version=beyond
# add "-march=core-avx2" to FFLAGS and CFLAGS in make.inc
nohup make pw ph pp pwcond all_currents  kcw neb cp ld1 tddfpt xspectra hp epw w90 yambo  d3q &
nohup make  pwcond all_currents  kcw neb cp ld1 tddfpt xspectra hp epw w90 yambo  d3q &
make install

1.6 QE 7.1 + ELPA 2021.11 + gcc7 + OpenMPI 3.1.6 + MKL 2015(test)

cd elpa-2021.11.001
FC=mpiifort  CC=icc ./configure --prefix=/home/ywmu/backup/Src/comlib/elpa FCFLAGS="-O3" CFLAGS="-O3" --enable-option-checking=fatal  SCALAPACK_LDFLAGS="-L$MKL_HOME/lib/intel64 -lmkl_scalapack_lp64 -lmkl_intel_lp64 -lmkl_sequential -lmkl_core -lmkl_blacs_intelmpi_lp64 -lpthread -lm -Wl,-rpath,$MKL_HOME/lib/intel64" SCALAPACK_FCFLAGS="-L$MKL_HOME/lib/intel64 -lmkl_scalapack_lp64 -lmkl_intel_lp64 -lmkl_sequential -lmkl_core -lmkl_blacs_intelmpi_lp64 -lpthread -I$MKL_HOME/include/intel64/lp64" --enable-avx2
make
make install 

cd q-e-qe-7.1
export FFTW_INCLUDE=/home/ywmu/backup/Src/comlib/fftw-3.3.8.g7/include 
export ELPAROOT=/home/ywmu/backup/Src/comlib/elpa
export ELPAVER=elpa-2021.11.001

./configure FC=gfortran F90=gfortran F77=gfortran CC=gcc MPIF90=mpif90 FFT_LIBS="/home/ywmu/backup/Src/comlib/fftw-3.3.8.g7/lib/libfftw3.a"  BLAS_LIBS="/home/ywmu/backup/Src/comlib/gnu/libopenblas.a" --with-scalapack=yes SCALAPACK_LIBS="/home/ywmu/backup/Src/comlib/gnu/libscalapack.a" --prefix=/home/ywmu/backup/Src/qe/bin.ompi    -enable-parallel -enable-openmp --with-elpa-include="${ELPAROOT}/include/${ELPAVER}/modules" --with-elpa-lib="${ELPAROOT}/lib/libelpa.a" --with-elpa-version=beyond
nohup make pw ph pp pwcond all_currents  kcw neb cp ld1 tddfpt xspectra hp epw w90 yambo  d3q &
make install

1.7 Wannier90 3.1

download wanni90 3.1 version
Load Envs for gcc and Openmpi

  tar -xvf v3.1.0.tar.gz
  cd wannier90-3.1.0
  cp config/make.inc.ifort make.inc
  ### Edit make.inc:
  COMMS  = mpi
  MPIF90 = mpif90
   LIBS = /home/ywmu/backup/Src/comlib/gnu/libopenblas.a
###  Edit Makefile
PREFIX ?= /home/ywmu/.soft/espresso/otherbin

make 
make intall

1.8 Install PGI Compiler

Download PGI Compiler from https://developer.nvidia.com/hpc-sdk

2 Xcrysden

2.1 Install On Linux

Refer to http://bbs.keinsci.com/thread-18599-1-1.html

Download shemi-shared version of XCrysDen 1.6.2 from http://sobereva.com/attach/564/xcrysden-1.6.2_sobereva.tar.gz
sudo yum install libGL-devel libGLU-devel libXmu-devel libXScrnSaver.x86_64
configure

tar -xvf xcrysden-1.6.2_sobereva.tar.gz
cd xcrysden-1.6.2
./scripts/xcConfigure.sh
... (set no crystal)

2.1.1 Defaut parameters

vi ~/.xcrysden/custom-definitions
set xcMisc(gif_encoder)   convert
set xcMisc(ImageMagick.convert) "/usr/bin/convert \
                             -quality 300 -border 3x3 -bordercolor black"

### change background color from back to white
set myParam(BACKGROUND)      {1.00 1.00 1.00}

2.1.2 Useful shortcuts

2.1.3 Main window

Ctrl+p: Print to image
Ctrl+s: save structure (e.g. as xsf)
Ctrl+t: save state and structure (e.g. as xsf)
y: Coordinate system
f: forces
F3/F4: Primitive/Convertional Cell
shift+a: atomic color
shift+c: cell color
shift+n: supercell view

2.1.4 Surface

t: Transparent Fermi Surface
c: Display first Brillouin zone
p: Crop Fermi surface to first BZ
shift+s: surface color
`shift+c: cell color

2.2 Install On Windows

2.2.1 Install Cygwin

Download Cygwin from http://www.cygwin.com/
Double click setup-x86_64.exe, install it step by step, Select Domestic sources （e.g., 163.com）, Install tcl-tk (8.6), tcl-togl, ImageMagick, xinit, mesa, libGLU1, libfftw3_3, libgfortran4 packages.

Set Environment for Xcrysden:

export  DISPLAY=:0
export PATH=$PATH:/usr/lib:/usr/lib/Togl2.0

2.2.2 Install Xcrysden

Download Xcrysden xcrysden-1.6.2-cygwin.tar.gz from http://www.xcrysden.org/, unzip it.
Run ./xcrysden from the its home directory.

2.3 Tips

The AdvGeom and Property modules are only active if the CRYSTAL package was installed, but the band path selection is available for all users.

3 Windows Gui for Quantum Espresso (QE)

	Burai	QuantumVITAS	QE on Device Studio
Input geometry format	xyz/POSCAR/xsf/pwscf/cif	Only pwscf(+g2g),not convenient	POSCAR/xyz/cif/hzw
Support	Relaxation/Cell Relaxation/Single Point/Band Structure/DOS/MD	Relaxation/Cell Relaxation/Single Point/Band Structure/DOS/MD/Transition State/TDDFT/Phonon	Relaxation/Cell Relaxation/Single Point/Band Structure/DOS
Modeler	Slab/Supercell/Translate	No(+Vesta/Avogadro2/PyMatGen/vaspkit)	Many functions from Device Studio
Optimization/MD Trace	Movie	Frames	No
Band Structure Output Analysis and Plot s	Bands, not convinent for K-path input	Fat band, Copy K-path from materialscloud	very convenient in Device Studio
On Development	no	yes	yes
Drawback:	supported tasks too few	input geometry is not convenient	MPI is Not supported. supported tasks too few

3.1 QE on Device Studio

Supported Tasks: Relaxation/Cell Relaxation/Single Point/Band Structure/DOS.
Drawback: MPI is Not supported. Analysis is too simple. The date for the generated images can not be exported.

3.1.1 More tasks

Creat a directory, such as D:\Temp\QEcmd
Copy mpi and qe from QuantumVITAS to D:\Temp\QEcmd
download busybox.exe to D:\Temp\QEcmd
Creat a QE_var.reg, and double click it. The content is as follows,

Windows Registry Editor Version 5.00

[HKEY_CLASSES_ROOT\Directory\Background\shell\RunQE]

[HKEY_CLASSES_ROOT\Directory\Background\shell\RunQE\command]
@="D:\\Temp\\QEcmd\\QE_init.bat"

Creat QE_init.bat,

@ECHO OFF
SET QEHOME=%~dp0
SET PATH=%QEHOME%mpi;%QEHOME%qe;%QEHOME%;%PATH%
busybox.exe sh

Go to the working directory, right click on the background, and click RunQE
Run tasks like,

export OMP_NUM_THREADS=1
mpiexec -np 2 pw.exe -in relax.in | tee  relax.out.pwscf

Open relax.out.pwscf with Avogadro2, and export it out as opt.cif.