Dear Z2pack team,
I have complied Wannier90 1.2 on top of VASP, and then I downloaded and installed z2pack like below:
(py35) [verali@farber input]$ python Python 3.5.2 |Anaconda 4.2.0 (64-bit)| (default, Jul 2 2016, 17:53:06) [GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux Type "help", "copyright", "credits" or "license" for more information.
import z2pack z2pack.__version__
'2.0.3'
I think this means I have successfully installed z2pack, right?
But when I run run.py script, it had the error:
File "run.py", line 14, in <module> Traceback (most recent call last): File "run.py", line 14, in <module> import z2pack ImportError: No module named z2pack Traceback (most recent call last): File "run.py", line 14, in <module>
Below is my job file:
vpkg_require intel/2015.3.187 vpkg_require openmpi/1.8.5
#$ -pe mpi 40 #$ -l m_mem_free=3G #$ -l h_cpu=09:00:00 ##$ -l exclusive=1
MY_EXE="/home/work/mtg/users/bin/vasp_ncl"
MY_EXE_ARGS=()
# # Uncomment to enable lots of additional information as OpenMPI executes # your job: # #SHOW_MPI_DEBUGGING=YES
## ## You should NOT need to change anything after this comment. ## OPENMPI_FLAGS="--display-map --mca btl ^openib --mca oob_tcp_if_exclude lo,ib0 --mca btl_tcp_if_exclude lo,ib0" if [ "${WANT_CPU_AFFINITY:-NO}" = "YES" ]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --bind-to core" fi if [ "${WANT_NPROC:-0}" -gt 0 ]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --np ${WANT_NPROC} --map-by node" fi if [ "${SHOW_MPI_DEBUGGING:-NO}" = "YES" ]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --debug-devel --debug-daemons --display-devel-map --display-devel-allocation --mca mca_verbose 1 --mca coll_base_verbose 1 --mca ras_base_verbose 1 --mca ras_gridengine_debug 1 --mca ras_gridengine_verbose 1 --mca btl_base_verbose 1 --mca mtl_base_verbose 1 --mca plm_base_verbose 1 --mca pls_rsh_debug 1" if [ "${WANT_CPU_AFFINITY:-NO}" = "YES" ]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --report-bindings" fi fi
echo "GridEngine parameters:" echo " mpirun = "`which mpirun` echo " nhosts = $NHOSTS” echo " nproc = $NSLOTS" echo " executable = $MY_EXE" echo " MPI flags = $OPENMPI_FLAGS" echo "-- begin OPENMPI run -- on " mpirun ${OPENMPI_FLAGS} python run.py echo "-- end OPENMPI run -- on “
Pls help to advise how can I solve this problem.
Thanks in advance.
Wei Li Ph.D student University of Delaware
Dear Wei Li,
The python version on which you installed Z2Pack (3.5.2) is probably not the same as you are using to run the script. Since it seems you are using a virtualenv (py35) when running Python from the command line, you should make sure the same virtualenv is also active in your job (i.e., "source py35/bin/activate" in your script).
Best regards,
Dominik Gresch
On 08.02.2017 10:31, Wei Li wrote:
Dear Z2pack team,
I have complied Wannier90 1.2 on top of VASP, and then I downloaded and installed z2pack like below:
(py35) [verali@farber input]$ python Python 3.5.2 |Anaconda 4.2.0 (64-bit)| (default, Jul 2 2016, 17:53:06) [GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux Type "help", "copyright", "credits" or "license" for more information.
import z2pack z2pack.__version__
'2.0.3'
I think this means I have successfully installed z2pack, right?
But when I run run.py script, it had the error:
File "run.py", line 14, in <module> Traceback (most recent call last): File "run.py", line 14, in <module> import z2pack ImportError: No module named z2pack Traceback (most recent call last): File "run.py", line 14, in <module>
*Below is my job file:*
vpkg_require intel/2015.3.187 vpkg_require openmpi/1.8.5
#$ -pe mpi 40 #$ -l m_mem_free=3G #$ -l h_cpu=09:00:00 ##$ -l exclusive=1
MY_EXE="/home/work/mtg/users/bin/vasp_ncl"
MY_EXE_ARGS=()
# # Uncomment to enable lots of additional information as OpenMPI executes # your job: # #SHOW_MPI_DEBUGGING=YES
## ## You should NOT need to change anything after this comment. ## OPENMPI_FLAGS="--display-map --mca btl ^openib --mca oob_tcp_if_exclude lo,ib0 --mca btl_tcp_if_exclude lo,ib0" if [ "${WANT_CPU_AFFINITY:-NO}"= "YES"]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --bind-to core" fi if [ "${WANT_NPROC:-0}"-gt 0 ]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --np ${WANT_NPROC} --map-by node" fi if [ "${SHOW_MPI_DEBUGGING:-NO}"= "YES"]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --debug-devel --debug-daemons --display-devel-map --display-devel-allocation --mca mca_verbose 1 --mca coll_base_verbose 1 --mca ras_base_verbose 1 --mca ras_gridengine_debug 1 --mca ras_gridengine_verbose 1 --mca btl_base_verbose 1 --mca mtl_base_verbose 1 --mca plm_base_verbose 1 --mca pls_rsh_debug 1" if [ "${WANT_CPU_AFFINITY:-NO}"= "YES"]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --report-bindings" fi fi
echo "GridEngine parameters:" echo " mpirun = "`which mpirun` echo " nhosts = $NHOSTS” echo " nproc = $NSLOTS" echo " executable = $MY_EXE" echo " MPI flags = $OPENMPI_FLAGS" echo "-- begin OPENMPI run -- on " mpirun ${OPENMPI_FLAGS} python run.py echo "-- end OPENMPI run -- on “
Pls help to advise how can I solve this problem.
Thanks in advance.
Wei Li Ph.D student University of Delaware
Dear Dominik Gresch,
Thanks a lot.
I activated py35 and it works.
Do I need to change any parameters if I want to run the example given here: https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp
I run it directly and I got the attached output, I did not find any error in output file or log file, but the process is terminated.
Regards Wei Li
On Feb 8, 2017, at 4:53 AM, Dominik Gresch greschd@phys.ethz.ch wrote:
Dear Wei Li,
The python version on which you installed Z2Pack (3.5.2) is probably not the same as you are using to run the script. Since it seems you are using a virtualenv (py35) when running Python from the command line, you should make sure the same virtualenv is also active in your job (i.e., "source py35/bin/activate" in your script).
Best regards,
Dominik Gresch On 08.02.2017 10:31, Wei Li wrote:
Dear Z2pack team,
I have complied Wannier90 1.2 on top of VASP, and then I downloaded and installed z2pack like below:
(py35) [verali@farber input]$ python Python 3.5.2 |Anaconda 4.2.0 (64-bit)| (default, Jul 2 2016, 17:53:06) [GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux Type "help", "copyright", "credits" or "license" for more information.
import z2pack z2pack.__version__
'2.0.3'
I think this means I have successfully installed z2pack, right?
But when I run run.py script, it had the error:
File "run.py", line 14, in <module> Traceback (most recent call last): File "run.py", line 14, in <module> import z2pack ImportError: No module named z2pack Traceback (most recent call last): File "run.py", line 14, in <module>
Below is my job file:
vpkg_require intel/2015.3.187 vpkg_require openmpi/1.8.5
#$ -pe mpi 40 #$ -l m_mem_free=3G #$ -l h_cpu=09:00:00 ##$ -l exclusive=1
MY_EXE="/home/work/mtg/users/bin/vasp_ncl"
MY_EXE_ARGS=()
# # Uncomment to enable lots of additional information as OpenMPI executes # your job: # #SHOW_MPI_DEBUGGING=YES
## ## You should NOT need to change anything after this comment. ## OPENMPI_FLAGS="--display-map --mca btl ^openib --mca oob_tcp_if_exclude lo,ib0 --mca btl_tcp_if_exclude lo,ib0" if [ "${WANT_CPU_AFFINITY:-NO}" = "YES" ]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --bind-to core" fi if [ "${WANT_NPROC:-0}" -gt 0 ]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --np ${WANT_NPROC} --map-by node" fi if [ "${SHOW_MPI_DEBUGGING:-NO}" = "YES" ]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --debug-devel --debug-daemons --display-devel-map --display-devel-allocation --mca mca_verbose 1 --mca coll_base_verbose 1 --mca ras_base_verbose 1 --mca ras_gridengine_debug 1 --mca ras_gridengine_verbose 1 --mca btl_base_verbose 1 --mca mtl_base_verbose 1 --mca plm_base_verbose 1 --mca pls_rsh_debug 1" if [ "${WANT_CPU_AFFINITY:-NO}" = "YES" ]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --report-bindings" fi fi
echo "GridEngine parameters:" echo " mpirun = "`which mpirun` echo " nhosts = $NHOSTS” echo " nproc = $NSLOTS" echo " executable = $MY_EXE" echo " MPI flags = $OPENMPI_FLAGS" echo "-- begin OPENMPI run -- on " mpirun ${OPENMPI_FLAGS} python run.py echo "-- end OPENMPI run -- on “
Pls help to advise how can I solve this problem.
Thanks in advance.
Wei Li Ph.D student University of Delaware
Dear Wei Li,
It seems you are running multiple instances of the Python interpreter, due to the 'mpirun' command. There seems to be a bit of a race condition when they all try to create the 'results' folder.
In general, you shouldn't run Python itself with mpirun, only the first-principles command (command="mpirun YOUR_VASP_COMMAND" in the z2pack.fp.System options).
Best regards,
Dominik Gresch
On 08.02.2017 17:08, Wei Li wrote:
Dear Dominik Gresch,
Thanks a lot.
I activated py35 and it works.
Do I need to change any parameters if I want to run the example given here: https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp
I run it directly and I got the attached output, I did not find any error in output file or log file, but the process is terminated.
Regards Wei Li
On Feb 8, 2017, at 4:53 AM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
The python version on which you installed Z2Pack (3.5.2) is probably not the same as you are using to run the script. Since it seems you are using a virtualenv (py35) when running Python from the command line, you should make sure the same virtualenv is also active in your job (i.e., "source py35/bin/activate" in your script).
Best regards,
Dominik Gresch
On 08.02.2017 10:31, Wei Li wrote:
Dear Z2pack team,
I have complied Wannier90 1.2 on top of VASP, and then I downloaded and installed z2pack like below:
(py35) [verali@farber input]$ python Python 3.5.2 |Anaconda 4.2.0 (64-bit)| (default, Jul 2 2016, 17:53:06) [GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux Type "help", "copyright", "credits" or "license" for more information.
import z2pack z2pack.__version__
'2.0.3'
I think this means I have successfully installed z2pack, right?
But when I run run.py script, it had the error:
File "run.py", line 14, in <module> Traceback (most recent call last): File "run.py", line 14, in <module> import z2pack ImportError: No module named z2pack Traceback (most recent call last): File "run.py", line 14, in <module>
*Below is my job file:*
vpkg_require intel/2015.3.187 vpkg_require openmpi/1.8.5
#$ -pe mpi 40 #$ -l m_mem_free=3G #$ -l h_cpu=09:00:00 ##$ -l exclusive=1
MY_EXE="/home/work/mtg/users/bin/vasp_ncl"
MY_EXE_ARGS=()
# # Uncomment to enable lots of additional information as OpenMPI executes # your job: # #SHOW_MPI_DEBUGGING=YES
## ## You should NOT need to change anything after this comment. ## OPENMPI_FLAGS="--display-map --mca btl ^openib --mca oob_tcp_if_exclude lo,ib0 --mca btl_tcp_if_exclude lo,ib0" if [ "${WANT_CPU_AFFINITY:-NO}"= "YES"]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --bind-to core" fi if [ "${WANT_NPROC:-0}"-gt 0 ]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --np ${WANT_NPROC} --map-by node" fi if [ "${SHOW_MPI_DEBUGGING:-NO}"= "YES"]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --debug-devel --debug-daemons --display-devel-map --display-devel-allocation --mca mca_verbose 1 --mca coll_base_verbose 1 --mca ras_base_verbose 1 --mca ras_gridengine_debug 1 --mca ras_gridengine_verbose 1 --mca btl_base_verbose 1 --mca mtl_base_verbose 1 --mca plm_base_verbose 1 --mca pls_rsh_debug 1" if [ "${WANT_CPU_AFFINITY:-NO}"= "YES"]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --report-bindings" fi fi
echo "GridEngine parameters:" echo " mpirun = "`which mpirun` echo " nhosts = $NHOSTS” echo " nproc = $NSLOTS" echo " executable = $MY_EXE" echo " MPI flags = $OPENMPI_FLAGS" echo "-- begin OPENMPI run -- on " mpirun ${OPENMPI_FLAGS} python run.py echo "-- end OPENMPI run -- on “
Pls help to advise how can I solve this problem.
Thanks in advance.
Wei Li Ph.D student University of Delaware
Dear Dominik,
Thanks for your reply. I solved this problem.
I generated the build file as following:
CHG CONTCAR EIGENVAL INCAR OSZICAR paral.qs.o889660 POSCAR REPORT wannier90.eig wannier90.win WAVECAR CHGCAR DOSCAR IBZKPT KPOINTS OUTCAR PCDAT POTCAR vasprun.xml wannier90.mmn wannier90.wout XDATCAR
There is no error.
which script do I need to read the .mmn file to plot the result? So far, the result file is empty.
Regards Vera
On Feb 8, 2017, at 11:16 AM, Dominik Gresch greschd@phys.ethz.ch wrote:
Dear Wei Li,
It seems you are running multiple instances of the Python interpreter, due to the 'mpirun' command. There seems to be a bit of a race condition when they all try to create the 'results' folder.
In general, you shouldn't run Python itself with mpirun, only the first-principles command (command="mpirun YOUR_VASP_COMMAND" in the z2pack.fp.System options). Best regards,
Dominik Gresch
On 08.02.2017 17:08, Wei Li wrote:
Dear Dominik Gresch,
Thanks a lot.
I activated py35 and it works.
Do I need to change any parameters if I want to run the example given here: https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp
I run it directly and I got the attached output, I did not find any error in output file or log file, but the process is terminated.
Regards Wei Li
On Feb 8, 2017, at 4:53 AM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
The python version on which you installed Z2Pack (3.5.2) is probably not the same as you are using to run the script. Since it seems you are using a virtualenv (py35) when running Python from the command line, you should make sure the same virtualenv is also active in your job (i.e., "source py35/bin/activate" in your script).
Best regards,
Dominik Gresch On 08.02.2017 10:31, Wei Li wrote:
Dear Z2pack team,
I have complied Wannier90 1.2 on top of VASP, and then I downloaded and installed z2pack like below:
(py35) [verali@farber input]$ python Python 3.5.2 |Anaconda 4.2.0 (64-bit)| (default, Jul 2 2016, 17:53:06) [GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux Type "help", "copyright", "credits" or "license" for more information.
> import z2pack > z2pack.__version__
'2.0.3'
>
I think this means I have successfully installed z2pack, right?
But when I run run.py script, it had the error:
File "run.py", line 14, in <module> Traceback (most recent call last): File "run.py", line 14, in <module> import z2pack ImportError: No module named z2pack Traceback (most recent call last): File "run.py", line 14, in <module>
Below is my job file:
vpkg_require intel/2015.3.187 vpkg_require openmpi/1.8.5
#$ -pe mpi 40 #$ -l m_mem_free=3G #$ -l h_cpu=09:00:00 ##$ -l exclusive=1
MY_EXE="/home/work/mtg/users/bin/vasp_ncl"
MY_EXE_ARGS=()
# # Uncomment to enable lots of additional information as OpenMPI executes # your job: # #SHOW_MPI_DEBUGGING=YES
## ## You should NOT need to change anything after this comment. ## OPENMPI_FLAGS="--display-map --mca btl ^openib --mca oob_tcp_if_exclude lo,ib0 --mca btl_tcp_if_exclude lo,ib0" if [ "${WANT_CPU_AFFINITY:-NO}" = "YES" ]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --bind-to core" fi if [ "${WANT_NPROC:-0}" -gt 0 ]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --np ${WANT_NPROC} --map-by node" fi if [ "${SHOW_MPI_DEBUGGING:-NO}" = "YES" ]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --debug-devel --debug-daemons --display-devel-map --display-devel-allocation --mca mca_verbose 1 --mca coll_base_verbose 1 --mca ras_base_verbose 1 --mca ras_gridengine_debug 1 --mca ras_gridengine_verbose 1 --mca btl_base_verbose 1 --mca mtl_base_verbose 1 --mca plm_base_verbose 1 --mca pls_rsh_debug 1" if [ "${WANT_CPU_AFFINITY:-NO}" = "YES" ]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --report-bindings" fi fi
echo "GridEngine parameters:" echo " mpirun = "`which mpirun` echo " nhosts = $NHOSTS” echo " nproc = $NSLOTS" echo " executable = $MY_EXE" echo " MPI flags = $OPENMPI_FLAGS" echo "-- begin OPENMPI run -- on " mpirun ${OPENMPI_FLAGS} python run.py echo "-- end OPENMPI run -- on “
Pls help to advise how can I solve this problem.
Thanks in advance.
Wei Li Ph.D student University of Delaware
Dear Dominik,
From what I understand, after calling the first principle code (in my case, it is VASP), the next step is to read the .mmn file in the top level of the build file. So I give the mmn path like below:
system = z2pack.fp.System( input_files=["input/CHGCAR", "input/INCAR", "input/POSCAR", "input/POTCAR", "input/wannier90.win" ], kpt_fct=z2pack.fp.kpoint.vasp, kpt_path="KPOINTS", command="qsub /home/work/mtg/users/verali/Bi_Z2/Z2/paral.qs" mmn_path="wannier90.mmn" )
But it does not work.
Could you advise how to read the mmn file?
Regards Wei Li
On Feb 10, 2017, at 2:59 PM, Wei Li verali@udel.edu wrote:
Dear Dominik,
Thanks for your reply. I solved this problem.
I generated the build file as following:
CHG CONTCAR EIGENVAL INCAR OSZICAR paral.qs.o889660 POSCAR REPORT wannier90.eig wannier90.win WAVECAR CHGCAR DOSCAR IBZKPT KPOINTS OUTCAR PCDAT POTCAR vasprun.xml wannier90.mmn wannier90.wout XDATCAR
There is no error.
which script do I need to read the .mmn file to plot the result? So far, the result file is empty.
Regards Vera
On Feb 8, 2017, at 11:16 AM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
It seems you are running multiple instances of the Python interpreter, due to the 'mpirun' command. There seems to be a bit of a race condition when they all try to create the 'results' folder.
In general, you shouldn't run Python itself with mpirun, only the first-principles command (command="mpirun YOUR_VASP_COMMAND" in the z2pack.fp.System options). Best regards,
Dominik Gresch
On 08.02.2017 17:08, Wei Li wrote:
Dear Dominik Gresch,
Thanks a lot.
I activated py35 and it works.
Do I need to change any parameters if I want to run the example given here: https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp
I run it directly and I got the attached output, I did not find any error in output file or log file, but the process is terminated.
Regards Wei Li
On Feb 8, 2017, at 4:53 AM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
The python version on which you installed Z2Pack (3.5.2) is probably not the same as you are using to run the script. Since it seems you are using a virtualenv (py35) when running Python from the command line, you should make sure the same virtualenv is also active in your job (i.e., "source py35/bin/activate" in your script).
Best regards,
Dominik Gresch On 08.02.2017 10:31, Wei Li wrote:
Dear Z2pack team,
I have complied Wannier90 1.2 on top of VASP, and then I downloaded and installed z2pack like below:
(py35) [verali@farber input]$ python Python 3.5.2 |Anaconda 4.2.0 (64-bit)| (default, Jul 2 2016, 17:53:06) [GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux Type "help", "copyright", "credits" or "license" for more information.
>> import z2pack >> z2pack.__version__
'2.0.3'
>>
I think this means I have successfully installed z2pack, right?
But when I run run.py script, it had the error:
File "run.py", line 14, in <module> Traceback (most recent call last): File "run.py", line 14, in <module> import z2pack ImportError: No module named z2pack Traceback (most recent call last): File "run.py", line 14, in <module>
Below is my job file:
vpkg_require intel/2015.3.187 vpkg_require openmpi/1.8.5
#$ -pe mpi 40 #$ -l m_mem_free=3G #$ -l h_cpu=09:00:00 ##$ -l exclusive=1
MY_EXE="/home/work/mtg/users/bin/vasp_ncl"
MY_EXE_ARGS=()
# # Uncomment to enable lots of additional information as OpenMPI executes # your job: # #SHOW_MPI_DEBUGGING=YES
## ## You should NOT need to change anything after this comment. ## OPENMPI_FLAGS="--display-map --mca btl ^openib --mca oob_tcp_if_exclude lo,ib0 --mca btl_tcp_if_exclude lo,ib0" if [ "${WANT_CPU_AFFINITY:-NO}" = "YES" ]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --bind-to core" fi if [ "${WANT_NPROC:-0}" -gt 0 ]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --np ${WANT_NPROC} --map-by node" fi if [ "${SHOW_MPI_DEBUGGING:-NO}" = "YES" ]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --debug-devel --debug-daemons --display-devel-map --display-devel-allocation --mca mca_verbose 1 --mca coll_base_verbose 1 --mca ras_base_verbose 1 --mca ras_gridengine_debug 1 --mca ras_gridengine_verbose 1 --mca btl_base_verbose 1 --mca mtl_base_verbose 1 --mca plm_base_verbose 1 --mca pls_rsh_debug 1" if [ "${WANT_CPU_AFFINITY:-NO}" = "YES" ]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --report-bindings" fi fi
echo "GridEngine parameters:" echo " mpirun = "`which mpirun` echo " nhosts = $NHOSTS” echo " nproc = $NSLOTS" echo " executable = $MY_EXE" echo " MPI flags = $OPENMPI_FLAGS" echo "-- begin OPENMPI run -- on " mpirun ${OPENMPI_FLAGS} python run.py echo "-- end OPENMPI run -- on “
Pls help to advise how can I solve this problem.
Thanks in advance.
Wei Li Ph.D student University of Delaware
Dear Wei Li,
If everything work correctly, Z2Pack should automatically call VASP (via the command given) and read the .mmn file after VASP has generated it. This is because Z2Pack needs to iteratively change the k-point path with which VASP is called and the overlap file is calculated.
In your case, it seems that you are submitting the VASP run to some scheduler -- this means the command returns immediately, before the .mmn file has been generated. As a result, Z2Pack tries to read the file, but it does not exist yet. You can solve this problem by submitting instead the Python process running Z2Pack, and then running VASP within that same job.
Best regards,
Dominik
On 11.02.2017 20:40, Wei Li wrote:
Dear Dominik,
From what I understand, after calling the first principle code (in my case, it is VASP), the next step is to read the .mmn file in the top level of the build file. So I give the mmn path like below:
system = z2pack.fp.System( input_files=["input/CHGCAR", "input/INCAR", "input/POSCAR", "input/POTCAR", "input/wannier90.win"], kpt_fct=z2pack.fp.kpoint.vasp, kpt_path="KPOINTS", command="qsub /home/work/mtg/users/verali/Bi_Z2/Z2/paral.qs" mmn_path="wannier90.mmn" )
But it does not work.
Could you advise how to read the mmn file?
Regards Wei Li
On Feb 10, 2017, at 2:59 PM, Wei Li <verali@udel.edu mailto:verali@udel.edu> wrote:
Dear Dominik,
Thanks for your reply. I solved this problem.
I generated the build file as following:
CHG CONTCAR EIGENVAL INCAR OSZICAR paral.qs.o889660 POSCAR REPORT wannier90.eig wannier90.win WAVECAR CHGCAR DOSCAR IBZKPT KPOINTS OUTCAR PCDAT POTCAR vasprun.xml wannier90.mmn wannier90.wout XDATCAR
There is no error.
which script do I need to read the .mmn file to plot the result? So far, the result file is empty.
Regards Vera
On Feb 8, 2017, at 11:16 AM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
It seems you are running multiple instances of the Python interpreter, due to the 'mpirun' command. There seems to be a bit of a race condition when they all try to create the 'results' folder.
In general, you shouldn't run Python itself with mpirun, only the first-principles command (command="mpirun YOUR_VASP_COMMAND" in the z2pack.fp.System options).
Best regards,
Dominik Gresch
On 08.02.2017 17:08, Wei Li wrote:
Dear Dominik Gresch,
Thanks a lot.
I activated py35 and it works.
Do I need to change any parameters if I want to run the example given here: https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp
I run it directly and I got the attached output, I did not find any error in output file or log file, but the process is terminated.
Regards Wei Li
On Feb 8, 2017, at 4:53 AM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
The python version on which you installed Z2Pack (3.5.2) is probably not the same as you are using to run the script. Since it seems you are using a virtualenv (py35) when running Python from the command line, you should make sure the same virtualenv is also active in your job (i.e., "source py35/bin/activate" in your script).
Best regards,
Dominik Gresch
On 08.02.2017 10:31, Wei Li wrote:
Dear Z2pack team,
I have complied Wannier90 1.2 on top of VASP, and then I downloaded and installed z2pack like below:
(py35) [verali@farber input]$ python Python 3.5.2 |Anaconda 4.2.0 (64-bit)| (default, Jul 2 2016, 17:53:06) [GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux Type "help", "copyright", "credits" or "license" for more information. >>> import z2pack >>> z2pack.__version__ '2.0.3' >>>
I think this means I have successfully installed z2pack, right?
But when I run run.py script, it had the error:
File "run.py", line 14, in <module> Traceback (most recent call last): File "run.py", line 14, in <module> import z2pack ImportError: No module named z2pack Traceback (most recent call last): File "run.py", line 14, in <module>
*Below is my job file:*
vpkg_require intel/2015.3.187 vpkg_require openmpi/1.8.5
#$ -pe mpi 40 #$ -l m_mem_free=3G #$ -l h_cpu=09:00:00 ##$ -l exclusive=1
MY_EXE="/home/work/mtg/users/bin/vasp_ncl"
MY_EXE_ARGS=()
# # Uncomment to enable lots of additional information as OpenMPI executes # your job: # #SHOW_MPI_DEBUGGING=YES
## ## You should NOT need to change anything after this comment. ## OPENMPI_FLAGS="--display-map --mca btl ^openib --mca oob_tcp_if_exclude lo,ib0 --mca btl_tcp_if_exclude lo,ib0" if [ "${WANT_CPU_AFFINITY:-NO}"= "YES"]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --bind-to core" fi if [ "${WANT_NPROC:-0}"-gt 0 ]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --np ${WANT_NPROC} --map-by node" fi if [ "${SHOW_MPI_DEBUGGING:-NO}"= "YES"]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --debug-devel --debug-daemons --display-devel-map --display-devel-allocation --mca mca_verbose 1 --mca coll_base_verbose 1 --mca ras_base_verbose 1 --mca ras_gridengine_debug 1 --mca ras_gridengine_verbose 1 --mca btl_base_verbose 1 --mca mtl_base_verbose 1 --mca plm_base_verbose 1 --mca pls_rsh_debug 1" if [ "${WANT_CPU_AFFINITY:-NO}"= "YES"]; then OPENMPI_FLAGS="${OPENMPI_FLAGS} --report-bindings" fi fi
echo "GridEngine parameters:" echo " mpirun = "`which mpirun` echo " nhosts = $NHOSTS” echo " nproc = $NSLOTS" echo " executable = $MY_EXE" echo " MPI flags = $OPENMPI_FLAGS" echo "-- begin OPENMPI run -- on " mpirun ${OPENMPI_FLAGS} python run.py echo "-- end OPENMPI run -- on “
Pls help to advise how can I solve this problem.
Thanks in advance.
Wei Li Ph.D student University of Delaware
Hi Dominik,
Thanks, it runs well now.
For the given example (Bi) calculation, I need to set up the search_shells number. What’s the value you recommend? I have increased to 30 from default value 12, it still can not satisfy the B1 condition:
Unable to satisfy B1 with any of the first 30 shells Your cell might be very long, or you may have an irregular MP grid Try increasing the parameter search_shells in the win file (default=12)
Or do I need to check other parameters?
Here is my input:
num_wann = 10 num_bands = 10 spinors : true search_shells=30 num_iter 0 exclude_bands 11-15
Regards Wei Li
On Feb 11, 2017, at 4:00 PM, Dominik Gresch greschd@phys.ethz.ch wrote:
Dear Wei Li,
If everything work correctly, Z2Pack should automatically call VASP (via the command given) and read the .mmn file after VASP has generated it. This is because Z2Pack needs to iteratively change the k-point path with which VASP is called and the overlap file is calculated.
In your case, it seems that you are submitting the VASP run to some scheduler -- this means the command returns immediately, before the .mmn file has been generated. As a result, Z2Pack tries to read the file, but it does not exist yet. You can solve this problem by submitting instead the Python process running Z2Pack, and then running VASP within that same job.
Best regards,
Dominik
On 11.02.2017 20:40, Wei Li wrote:
Dear Dominik,
From what I understand, after calling the first principle code (in my case, it is VASP), the next step is to read the .mmn file in the top level of the build file. So I give the mmn path like below:
system = z2pack.fp.System( input_files=["input/CHGCAR", "input/INCAR", "input/POSCAR", "input/POTCAR", "input/wannier90.win" ], kpt_fct=z2pack.fp.kpoint.vasp, kpt_path="KPOINTS", command="qsub /home/work/mtg/users/verali/Bi_Z2/Z2/paral.qs" mmn_path="wannier90.mmn" )
But it does not work.
Could you advise how to read the mmn file?
Regards Wei Li
On Feb 10, 2017, at 2:59 PM, Wei Li <verali@udel.edu mailto:verali@udel.edu> wrote:
Dear Dominik,
Thanks for your reply. I solved this problem.
I generated the build file as following:
CHG CONTCAR EIGENVAL INCAR OSZICAR paral.qs.o889660 POSCAR REPORT wannier90.eig wannier90.win WAVECAR CHGCAR DOSCAR IBZKPT KPOINTS OUTCAR PCDAT POTCAR vasprun.xml wannier90.mmn wannier90.wout XDATCAR
There is no error.
which script do I need to read the .mmn file to plot the result? So far, the result file is empty.
Regards Vera
On Feb 8, 2017, at 11:16 AM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
It seems you are running multiple instances of the Python interpreter, due to the 'mpirun' command. There seems to be a bit of a race condition when they all try to create the 'results' folder.
In general, you shouldn't run Python itself with mpirun, only the first-principles command (command="mpirun YOUR_VASP_COMMAND" in the z2pack.fp.System options). Best regards,
Dominik Gresch
On 08.02.2017 17:08, Wei Li wrote:
Dear Dominik Gresch,
Thanks a lot.
I activated py35 and it works.
Do I need to change any parameters if I want to run the example given here: https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp
I run it directly and I got the attached output, I did not find any error in output file or log file, but the process is terminated.
Regards Wei Li
On Feb 8, 2017, at 4:53 AM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
The python version on which you installed Z2Pack (3.5.2) is probably not the same as you are using to run the script. Since it seems you are using a virtualenv (py35) when running Python from the command line, you should make sure the same virtualenv is also active in your job (i.e., "source py35/bin/activate" in your script).
Best regards,
Dominik Gresch On 08.02.2017 10:31, Wei Li wrote: > Dear Z2pack team, > > I have complied Wannier90 1.2 on top of VASP, and then I downloaded and installed z2pack like below: > > (py35) [verali@farber input]$ python > Python 3.5.2 |Anaconda 4.2.0 (64-bit)| (default, Jul 2 2016, 17:53:06) > [GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux > Type "help", "copyright", "credits" or "license" for more information. > >>> import z2pack > >>> z2pack.__version__ > '2.0.3' > >>> > > I think this means I have successfully installed z2pack, right? > > But when I run run.py script, it had the error: > > File "run.py", line 14, in <module> > Traceback (most recent call last): > File "run.py", line 14, in <module> > import z2pack > ImportError: No module named z2pack > Traceback (most recent call last): > File "run.py", line 14, in <module> > > > Below is my job file: > > > > vpkg_require intel/2015.3.187 > vpkg_require openmpi/1.8.5 > > #$ -pe mpi 40 > #$ -l m_mem_free=3G > #$ -l h_cpu=09:00:00 > ##$ -l exclusive=1 > > MY_EXE="/home/work/mtg/users/bin/vasp_ncl" > > MY_EXE_ARGS=() > > # > # Uncomment to enable lots of additional information as OpenMPI executes > # your job: > # > #SHOW_MPI_DEBUGGING=YES > > ## > ## You should NOT need to change anything after this comment. > ## > OPENMPI_FLAGS="--display-map --mca btl ^openib --mca oob_tcp_if_exclude lo,ib0 --mca btl_tcp_if_exclude lo,ib0" > if [ "${WANT_CPU_AFFINITY:-NO}" = "YES" ]; then > OPENMPI_FLAGS="${OPENMPI_FLAGS} --bind-to core" > fi > if [ "${WANT_NPROC:-0}" -gt 0 ]; then > OPENMPI_FLAGS="${OPENMPI_FLAGS} --np ${WANT_NPROC} --map-by node" > fi > if [ "${SHOW_MPI_DEBUGGING:-NO}" = "YES" ]; then > OPENMPI_FLAGS="${OPENMPI_FLAGS} --debug-devel --debug-daemons --display-devel-map --display-devel-allocation --mca mca_verbose 1 --mca coll_base_verbose 1 --mca ras_base_verbose 1 --mca ras_gridengine_debug 1 --mca ras_gridengine_verbose 1 --mca btl_base_verbose 1 --mca mtl_base_verbose 1 --mca plm_base_verbose 1 --mca pls_rsh_debug 1" > if [ "${WANT_CPU_AFFINITY:-NO}" = "YES" ]; then > OPENMPI_FLAGS="${OPENMPI_FLAGS} --report-bindings" > fi > fi > > echo "GridEngine parameters:" > echo " mpirun = "`which mpirun` > echo " nhosts = $NHOSTS” > echo " nproc = $NSLOTS" > echo " executable = $MY_EXE" > echo " MPI flags = $OPENMPI_FLAGS" > echo "-- begin OPENMPI run -- on " > mpirun ${OPENMPI_FLAGS} python run.py > echo "-- end OPENMPI run -- on “ > > Pls help to advise how can I solve this problem. > > Thanks in advance. > > Wei Li > Ph.D student > University of Delaware > >
Dear Wei Li,
The Bi example should run with the supplied input files. However, you need to install a modified version of Wannier90, as explained in the installation instructions: http://z2pack.ethz.ch/doc/2.0/tutorial_installation.html#setting-up-first-pr...
For Wannier90 2.1 or newer this is no longer necessary because we added a specific interface to run with Z2Pack, but unfortunately VASP only supports Wannier90 version 1.x.
Best regards,
Dominik Gresch
On 13.02.2017 23:40, Wei Li wrote:
Hi Dominik,
Thanks, it runs well now.
For the given example (Bi) calculation, I need to set up the search_shells number. What’s the value you recommend? I have increased to 30 from default value 12, it still can not satisfy the B1 condition:
Unable to satisfy B1 with any of the first 30 shells Your cell might be very long, or you may have an irregular MP grid Try increasing the parameter search_shells in the win file (default=12)
Or do I need to check other parameters?
Here is my input:
num_wann = 10 num_bands = 10 spinors : true search_shells=30 num_iter 0 exclude_bands 11-15
Regards Wei Li
On Feb 11, 2017, at 4:00 PM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
If everything work correctly, Z2Pack should automatically call VASP (via the command given) and read the .mmn file after VASP has generated it. This is because Z2Pack needs to iteratively change the k-point path with which VASP is called and the overlap file is calculated.
In your case, it seems that you are submitting the VASP run to some scheduler -- this means the command returns immediately, before the .mmn file has been generated. As a result, Z2Pack tries to read the file, but it does not exist yet. You can solve this problem by submitting instead the Python process running Z2Pack, and then running VASP within that same job.
Best regards,
Dominik
On 11.02.2017 20:40, Wei Li wrote:
Dear Dominik,
From what I understand, after calling the first principle code (in my case, it is VASP), the next step is to read the .mmn file in the top level of the build file. So I give the mmn path like below:
system = z2pack.fp.System( input_files=["input/CHGCAR", "input/INCAR", "input/POSCAR", "input/POTCAR", "input/wannier90.win"], kpt_fct=z2pack.fp.kpoint.vasp, kpt_path="KPOINTS", command="qsub /home/work/mtg/users/verali/Bi_Z2/Z2/paral.qs" mmn_path="wannier90.mmn" )
But it does not work.
Could you advise how to read the mmn file?
Regards Wei Li
On Feb 10, 2017, at 2:59 PM, Wei Li <verali@udel.edu mailto:verali@udel.edu> wrote:
Dear Dominik,
Thanks for your reply. I solved this problem.
I generated the build file as following:
CHG CONTCAR EIGENVAL INCAR OSZICAR paral.qs.o889660 POSCAR REPORT wannier90.eig wannier90.win WAVECAR CHGCAR DOSCAR IBZKPT KPOINTS OUTCAR PCDAT POTCAR vasprun.xml wannier90.mmn wannier90.wout XDATCAR
There is no error.
which script do I need to read the .mmn file to plot the result? So far, the result file is empty.
Regards Vera
On Feb 8, 2017, at 11:16 AM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
It seems you are running multiple instances of the Python interpreter, due to the 'mpirun' command. There seems to be a bit of a race condition when they all try to create the 'results' folder.
In general, you shouldn't run Python itself with mpirun, only the first-principles command (command="mpirun YOUR_VASP_COMMAND" in the z2pack.fp.System options).
Best regards,
Dominik Gresch
On 08.02.2017 17:08, Wei Li wrote:
Dear Dominik Gresch,
Thanks a lot.
I activated py35 and it works.
Do I need to change any parameters if I want to run the example given here: https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp
I run it directly and I got the attached output, I did not find any error in output file or log file, but the process is terminated.
Regards Wei Li
> On Feb 8, 2017, at 4:53 AM, Dominik Gresch <greschd@phys.ethz.ch > mailto:greschd@phys.ethz.ch> wrote: > > Dear Wei Li, > > The python version on which you installed Z2Pack (3.5.2) is > probably not the same as you are using to run the script. Since > it seems you are using a virtualenv (py35) when running Python > from the command line, you should make sure the same virtualenv > is also active in your job (i.e., "source py35/bin/activate" in > your script). > > Best regards, > > Dominik Gresch > > On 08.02.2017 10:31, Wei Li wrote: >> Dear Z2pack team, >> >> I have complied Wannier90 1.2 on top of VASP, and then I >> downloaded and installed z2pack like below: >> >> (py35) [verali@farber input]$ python >> Python 3.5.2 |Anaconda 4.2.0 (64-bit)| (default, Jul 2 2016, >> 17:53:06) >> [GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux >> Type "help", "copyright", "credits" or "license" for more >> information. >> >>> import z2pack >> >>> z2pack.__version__ >> '2.0.3' >> >>> >> >> I think this means I have successfully installed z2pack, right? >> >> But when I run run.py script, it had the error: >> >> File "run.py", line 14, in <module> >> Traceback (most recent call last): >> File "run.py", line 14, in <module> >> import z2pack >> ImportError: No module named z2pack >> Traceback (most recent call last): >> File "run.py", line 14, in <module> >> >> >> *Below is my job file:* >> >> >> >> vpkg_require intel/2015.3.187 >> vpkg_require openmpi/1.8.5 >> >> #$ -pe mpi 40 >> #$ -l m_mem_free=3G >> #$ -l h_cpu=09:00:00 >> ##$ -l exclusive=1 >> >> MY_EXE="/home/work/mtg/users/bin/vasp_ncl" >> >> MY_EXE_ARGS=() >> >> # >> # Uncomment to enable lots of additional information as OpenMPI >> executes >> # your job: >> # >> #SHOW_MPI_DEBUGGING=YES >> >> ## >> ## You should NOT need to change anything after this comment. >> ## >> OPENMPI_FLAGS="--display-map --mca btl ^openib --mca >> oob_tcp_if_exclude lo,ib0 --mca btl_tcp_if_exclude lo,ib0" >> if [ "${WANT_CPU_AFFINITY:-NO}"= "YES"]; then >> OPENMPI_FLAGS="${OPENMPI_FLAGS} --bind-to core" >> fi >> if [ "${WANT_NPROC:-0}"-gt 0 ]; then >> OPENMPI_FLAGS="${OPENMPI_FLAGS} --np ${WANT_NPROC} --map-by node" >> fi >> if [ "${SHOW_MPI_DEBUGGING:-NO}"= "YES"]; then >> OPENMPI_FLAGS="${OPENMPI_FLAGS} --debug-devel --debug-daemons >> --display-devel-map --display-devel-allocation --mca >> mca_verbose 1 --mca coll_base_verbose 1 --mca ras_base_verbose >> 1 --mca ras_gridengine_debug 1 --mca ras_gridengine_verbose 1 >> --mca btl_base_verbose 1 --mca mtl_base_verbose 1 --mca >> plm_base_verbose 1 --mca pls_rsh_debug 1" >> if [ "${WANT_CPU_AFFINITY:-NO}"= "YES"]; then >> OPENMPI_FLAGS="${OPENMPI_FLAGS} --report-bindings" >> fi >> fi >> >> echo "GridEngine parameters:" >> echo " mpirun = "`which mpirun` >> echo " nhosts = $NHOSTS” >> echo " nproc = $NSLOTS" >> echo " executable = $MY_EXE" >> echo " MPI flags = $OPENMPI_FLAGS" >> echo "-- begin OPENMPI run -- on " >> mpirun ${OPENMPI_FLAGS} python run.py >> echo "-- end OPENMPI run -- on “ >> >> Pls help to advise how can I solve this problem. >> >> Thanks in advance. >> >> Wei Li >> Ph.D student >> University of Delaware >> >> >
Dear Dominik,
Thanks, I have complied wannier90 2.1 on top of VASP and I am doing the test currently.
I noticed that z2pack dynamically generate the KPOINTS files for WCC calculations. Everything runs well except that I am not sure about parameterizing the surface, which means that when the KPOINTS loop can end. So far my KPOINTS files reach to like below:
Automatic mesh 0 ! number of k-points = 0 ->automatic generation scheme Gamma ! generate a Gamma centered grid 1 1 17 ! subdivisions 0.0 0.03125 0.0 ! shift
Do I need to set a parameter to end the KPOINTS shifting, or I just need to waiting for the calculation completed as the parameters in the tutorial? So far, it realy consumes lots of computation time. Here is my results file information:
[wli9@br006 Z2]$ ls -artl results/ total 1026 -rw------- 1 wli9 mr4s90p 75853 Feb 16 16:13 res_0.p drwxr-xr-x 2 wli9 mr4s90p 33280 Feb 16 16:13 . drwxr-xr-x 6 wli9 mr4s90p 33280 Feb 16 16:13 ..
Regards Wei Li
On Feb 13, 2017, at 5:44 PM, Dominik Gresch greschd@phys.ethz.ch wrote:
Dear Wei Li,
The Bi example should run with the supplied input files. However, you need to install a modified version of Wannier90, as explained in the installation instructions: http://z2pack.ethz.ch/doc/2.0/tutorial_installation.html#setting-up-first-pr... http://z2pack.ethz.ch/doc/2.0/tutorial_installation.html#setting-up-first-principles-codes For Wannier90 2.1 or newer this is no longer necessary because we added a specific interface to run with Z2Pack, but unfortunately VASP only supports Wannier90 version 1.x.
Best regards,
Dominik Gresch
On 13.02.2017 23:40, Wei Li wrote:
Hi Dominik,
Thanks, it runs well now.
For the given example (Bi) calculation, I need to set up the search_shells number. What’s the value you recommend? I have increased to 30 from default value 12, it still can not satisfy the B1 condition:
Unable to satisfy B1 with any of the first 30 shells Your cell might be very long, or you may have an irregular MP grid Try increasing the parameter search_shells in the win file (default=12)
Or do I need to check other parameters?
Here is my input:
num_wann = 10 num_bands = 10 spinors : true search_shells=30 num_iter 0 exclude_bands 11-15
Regards Wei Li
On Feb 11, 2017, at 4:00 PM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
If everything work correctly, Z2Pack should automatically call VASP (via the command given) and read the .mmn file after VASP has generated it. This is because Z2Pack needs to iteratively change the k-point path with which VASP is called and the overlap file is calculated.
In your case, it seems that you are submitting the VASP run to some scheduler -- this means the command returns immediately, before the .mmn file has been generated. As a result, Z2Pack tries to read the file, but it does not exist yet. You can solve this problem by submitting instead the Python process running Z2Pack, and then running VASP within that same job.
Best regards,
Dominik
On 11.02.2017 20:40, Wei Li wrote:
Dear Dominik,
From what I understand, after calling the first principle code (in my case, it is VASP), the next step is to read the .mmn file in the top level of the build file. So I give the mmn path like below:
system = z2pack.fp.System( input_files=["input/CHGCAR", "input/INCAR", "input/POSCAR", "input/POTCAR", "input/wannier90.win" ], kpt_fct=z2pack.fp.kpoint.vasp, kpt_path="KPOINTS", command="qsub /home/work/mtg/users/verali/Bi_Z2/Z2/paral.qs" mmn_path="wannier90.mmn" )
But it does not work.
Could you advise how to read the mmn file?
Regards Wei Li
On Feb 10, 2017, at 2:59 PM, Wei Li <verali@udel.edu mailto:verali@udel.edu> wrote:
Dear Dominik,
Thanks for your reply. I solved this problem.
I generated the build file as following:
CHG CONTCAR EIGENVAL INCAR OSZICAR paral.qs.o889660 POSCAR REPORT wannier90.eig wannier90.win WAVECAR CHGCAR DOSCAR IBZKPT KPOINTS OUTCAR PCDAT POTCAR vasprun.xml wannier90.mmn wannier90.wout XDATCAR
There is no error.
which script do I need to read the .mmn file to plot the result? So far, the result file is empty.
Regards Vera
On Feb 8, 2017, at 11:16 AM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
It seems you are running multiple instances of the Python interpreter, due to the 'mpirun' command. There seems to be a bit of a race condition when they all try to create the 'results' folder.
In general, you shouldn't run Python itself with mpirun, only the first-principles command (command="mpirun YOUR_VASP_COMMAND" in the z2pack.fp.System options). Best regards,
Dominik Gresch
On 08.02.2017 17:08, Wei Li wrote: > Dear Dominik Gresch, > > Thanks a lot. > > I activated py35 and it works. > > Do I need to change any parameters if I want to run the example given here: > https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp > > I run it directly and I got the attached output, I did not find any error in output file or log file, but the process is terminated. > > > Regards > Wei Li > > >> On Feb 8, 2017, at 4:53 AM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote: >> >> Dear Wei Li, >> >> The python version on which you installed Z2Pack (3.5.2) is probably not the same as you are using to run the script. Since it seems you are using a virtualenv (py35) when running Python from the command line, you should make sure the same virtualenv is also active in your job (i.e., "source py35/bin/activate" in your script). >> >> Best regards, >> >> Dominik Gresch >> On 08.02.2017 10:31, Wei Li wrote: >>> Dear Z2pack team, >>> >>> I have complied Wannier90 1.2 on top of VASP, and then I downloaded and installed z2pack like below: >>> >>> (py35) [verali@farber input]$ python >>> Python 3.5.2 |Anaconda 4.2.0 (64-bit)| (default, Jul 2 2016, 17:53:06) >>> [GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux >>> Type "help", "copyright", "credits" or "license" for more information. >>> >>> import z2pack >>> >>> z2pack.__version__ >>> '2.0.3' >>> >>> >>> >>> I think this means I have successfully installed z2pack, right? >>> >>> But when I run run.py script, it had the error: >>> >>> File "run.py", line 14, in <module> >>> Traceback (most recent call last): >>> File "run.py", line 14, in <module> >>> import z2pack >>> ImportError: No module named z2pack >>> Traceback (most recent call last): >>> File "run.py", line 14, in <module> >>> >>> >>> Below is my job file: >>> >>> >>> >>> vpkg_require intel/2015.3.187 >>> vpkg_require openmpi/1.8.5 >>> >>> #$ -pe mpi 40 >>> #$ -l m_mem_free=3G >>> #$ -l h_cpu=09:00:00 >>> ##$ -l exclusive=1 >>> >>> MY_EXE="/home/work/mtg/users/bin/vasp_ncl" >>> >>> MY_EXE_ARGS=() >>> >>> # >>> # Uncomment to enable lots of additional information as OpenMPI executes >>> # your job: >>> # >>> #SHOW_MPI_DEBUGGING=YES >>> >>> ## >>> ## You should NOT need to change anything after this comment. >>> ## >>> OPENMPI_FLAGS="--display-map --mca btl ^openib --mca oob_tcp_if_exclude lo,ib0 --mca btl_tcp_if_exclude lo,ib0" >>> if [ "${WANT_CPU_AFFINITY:-NO}" = "YES" ]; then >>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --bind-to core" >>> fi >>> if [ "${WANT_NPROC:-0}" -gt 0 ]; then >>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --np ${WANT_NPROC} --map-by node" >>> fi >>> if [ "${SHOW_MPI_DEBUGGING:-NO}" = "YES" ]; then >>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --debug-devel --debug-daemons --display-devel-map --display-devel-allocation --mca mca_verbose 1 --mca coll_base_verbose 1 --mca ras_base_verbose 1 --mca ras_gridengine_debug 1 --mca ras_gridengine_verbose 1 --mca btl_base_verbose 1 --mca mtl_base_verbose 1 --mca plm_base_verbose 1 --mca pls_rsh_debug 1" >>> if [ "${WANT_CPU_AFFINITY:-NO}" = "YES" ]; then >>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --report-bindings" >>> fi >>> fi >>> >>> echo "GridEngine parameters:" >>> echo " mpirun = "`which mpirun` >>> echo " nhosts = $NHOSTS” >>> echo " nproc = $NSLOTS" >>> echo " executable = $MY_EXE" >>> echo " MPI flags = $OPENMPI_FLAGS" >>> echo "-- begin OPENMPI run -- on " >>> mpirun ${OPENMPI_FLAGS} python run.py >>> echo "-- end OPENMPI run -- on “ >>> >>> Pls help to advise how can I solve this problem. >>> >>> Thanks in advance. >>> >>> Wei Li >>> Ph.D student >>> University of Delaware >>> >>> >> >
Dear Wei Li,
The surface is parametrized as a function of two parameters, as described here: http://z2pack.ethz.ch/doc/2.0/tutorial_surface.html#basic-syntax
For VASP, it is important to note that the surface must be straight in the second parameter (that is, f(t1=const., t2) describes a straight line across the full BZ, parallel to one of the reciprocal axes). This is again due to a limitation in the interface between Wannier90 and VASP, where only automatic mesh generation is possible.
For your second question about the progress of the calculation: To see the progress of your calculation, it is much more helpful to look at the output of Z2Pack itself instead of the build folder or the result file. Z2Pack reports which lines it is calculating, and what the status of the convergence is.
You can improve the performance for example if you find that the iteration in a single line never converges before a specific value, by adjusting the 'iterator' parameter accordingly. The main reason it can sometimes take a long time is because of the adaptive convergence criteria. This is a tradeoff we made on purpose, because it significantly improves the reliability of the results. Of course you can turn them off, but I would strongly advise against this. All in all, performance was not so much the focus in developing Z2Pack as reliability and flexibility in using it for different purposes.
Best regards,
Dominik
On 16.02.2017 22:15, Wei Li wrote:
Dear Dominik,
Thanks, I have complied wannier90 2.1 on top of VASP and I am doing the test currently.
I noticed that z2pack dynamically generate the KPOINTS files for WCC calculations. Everything runs well except that I am not sure about parameterizing the surface, which means that when the KPOINTS loop can end. So far my KPOINTS files reach to like below:
Automatic mesh 0 ! number of k-points = 0 ->automatic generation scheme Gamma ! generate a Gamma centered grid 1 1 17 ! subdivisions 0.0 0.03125 0.0! shift
Do I need to set a parameter to end the KPOINTS shifting, or I just need to waiting for the calculation completed as the parameters in the tutorial? So far, it realy consumes lots of computation time. Here is my results file information:
[wli9@br006 Z2]$ ls -artl results/ total 1026 -rw------- 1 wli9 mr4s90p 75853 Feb 16 16:13 res_0.p drwxr-xr-x 2 wli9 mr4s90p 33280 Feb 16 16:13 . drwxr-xr-x 6 wli9 mr4s90p 33280 Feb 16 16:13 ..
Regards Wei Li
On Feb 13, 2017, at 5:44 PM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
The Bi example should run with the supplied input files. However, you need to install a modified version of Wannier90, as explained in the installation instructions: http://z2pack.ethz.ch/doc/2.0/tutorial_installation.html#setting-up-first-pr...
For Wannier90 2.1 or newer this is no longer necessary because we added a specific interface to run with Z2Pack, but unfortunately VASP only supports Wannier90 version 1.x.
Best regards,
Dominik Gresch
On 13.02.2017 23:40, Wei Li wrote:
Hi Dominik,
Thanks, it runs well now.
For the given example (Bi) calculation, I need to set up the search_shells number. What’s the value you recommend? I have increased to 30 from default value 12, it still can not satisfy the B1 condition:
Unable to satisfy B1 with any of the first 30 shells Your cell might be very long, or you may have an irregular MP grid Try increasing the parameter search_shells in the win file (default=12)
Or do I need to check other parameters?
Here is my input:
num_wann = 10 num_bands = 10 spinors : true search_shells=30 num_iter 0 exclude_bands 11-15
Regards Wei Li
On Feb 11, 2017, at 4:00 PM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
If everything work correctly, Z2Pack should automatically call VASP (via the command given) and read the .mmn file after VASP has generated it. This is because Z2Pack needs to iteratively change the k-point path with which VASP is called and the overlap file is calculated.
In your case, it seems that you are submitting the VASP run to some scheduler -- this means the command returns immediately, before the .mmn file has been generated. As a result, Z2Pack tries to read the file, but it does not exist yet. You can solve this problem by submitting instead the Python process running Z2Pack, and then running VASP within that same job.
Best regards,
Dominik
On 11.02.2017 20:40, Wei Li wrote:
Dear Dominik,
From what I understand, after calling the first principle code (in my case, it is VASP), the next step is to read the .mmn file in the top level of the build file. So I give the mmn path like below:
system = z2pack.fp.System( input_files=["input/CHGCAR", "input/INCAR", "input/POSCAR", "input/POTCAR", "input/wannier90.win"], kpt_fct=z2pack.fp.kpoint.vasp, kpt_path="KPOINTS", command="qsub /home/work/mtg/users/verali/Bi_Z2/Z2/paral.qs" mmn_path="wannier90.mmn" )
But it does not work.
Could you advise how to read the mmn file?
Regards Wei Li
On Feb 10, 2017, at 2:59 PM, Wei Li <verali@udel.edu mailto:verali@udel.edu> wrote:
Dear Dominik,
Thanks for your reply. I solved this problem.
I generated the build file as following:
CHG CONTCAR EIGENVAL INCAR OSZICAR paral.qs.o889660 POSCAR REPORT wannier90.eig wannier90.win WAVECAR CHGCAR DOSCAR IBZKPT KPOINTS OUTCAR PCDAT POTCAR vasprun.xml wannier90.mmn wannier90.wout XDATCAR
There is no error.
which script do I need to read the .mmn file to plot the result? So far, the result file is empty.
Regards Vera
> On Feb 8, 2017, at 11:16 AM, Dominik Gresch > <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote: > > Dear Wei Li, > > It seems you are running multiple instances of the Python > interpreter, due to the 'mpirun' command. There seems to be a > bit of a race condition when they all try to create the > 'results' folder. > > In general, you shouldn't run Python itself with mpirun, only > the first-principles command (command="mpirun YOUR_VASP_COMMAND" > in the z2pack.fp.System options). > > Best regards, > > Dominik Gresch > > > On 08.02.2017 17:08, Wei Li wrote: >> Dear Dominik Gresch, >> >> Thanks a lot. >> >> I activated py35 and it works. >> >> Do I need to change any parameters if I want to run the example >> given here: >> https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp >> >> I run it directly and I got the attached output, I did not find >> any error in output file or log file, but the process is >> terminated. >> >> >> Regards >> Wei Li >> >> >>> On Feb 8, 2017, at 4:53 AM, Dominik Gresch >>> <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote: >>> >>> Dear Wei Li, >>> >>> The python version on which you installed Z2Pack (3.5.2) is >>> probably not the same as you are using to run the script. >>> Since it seems you are using a virtualenv (py35) when running >>> Python from the command line, you should make sure the same >>> virtualenv is also active in your job (i.e., "source >>> py35/bin/activate" in your script). >>> >>> Best regards, >>> >>> Dominik Gresch >>> >>> On 08.02.2017 10:31, Wei Li wrote: >>>> Dear Z2pack team, >>>> >>>> I have complied Wannier90 1.2 on top of VASP, and then I >>>> downloaded and installed z2pack like below: >>>> >>>> (py35) [verali@farber input]$ python >>>> Python 3.5.2 |Anaconda 4.2.0 (64-bit)| (default, Jul 2 2016, >>>> 17:53:06) >>>> [GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux >>>> Type "help", "copyright", "credits" or "license" for more >>>> information. >>>> >>> import z2pack >>>> >>> z2pack.__version__ >>>> '2.0.3' >>>> >>> >>>> >>>> I think this means I have successfully installed z2pack, right? >>>> >>>> But when I run run.py script, it had the error: >>>> >>>> File "run.py", line 14, in <module> >>>> Traceback (most recent call last): >>>> File "run.py", line 14, in <module> >>>> import z2pack >>>> ImportError: No module named z2pack >>>> Traceback (most recent call last): >>>> File "run.py", line 14, in <module> >>>> >>>> >>>> *Below is my job file:* >>>> >>>> >>>> >>>> vpkg_require intel/2015.3.187 >>>> vpkg_require openmpi/1.8.5 >>>> >>>> #$ -pe mpi 40 >>>> #$ -l m_mem_free=3G >>>> #$ -l h_cpu=09:00:00 >>>> ##$ -l exclusive=1 >>>> >>>> MY_EXE="/home/work/mtg/users/bin/vasp_ncl" >>>> >>>> MY_EXE_ARGS=() >>>> >>>> # >>>> # Uncomment to enable lots of additional information as >>>> OpenMPI executes >>>> # your job: >>>> # >>>> #SHOW_MPI_DEBUGGING=YES >>>> >>>> ## >>>> ## You should NOT need to change anything after this comment. >>>> ## >>>> OPENMPI_FLAGS="--display-map --mca btl ^openib --mca >>>> oob_tcp_if_exclude lo,ib0 --mca btl_tcp_if_exclude lo,ib0" >>>> if [ "${WANT_CPU_AFFINITY:-NO}"= "YES"]; then >>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --bind-to core" >>>> fi >>>> if [ "${WANT_NPROC:-0}"-gt 0 ]; then >>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --np ${WANT_NPROC} --map-by >>>> node" >>>> fi >>>> if [ "${SHOW_MPI_DEBUGGING:-NO}"= "YES"]; then >>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --debug-devel >>>> --debug-daemons --display-devel-map >>>> --display-devel-allocation --mca mca_verbose 1 --mca >>>> coll_base_verbose 1 --mca ras_base_verbose 1 --mca >>>> ras_gridengine_debug 1 --mca ras_gridengine_verbose 1 --mca >>>> btl_base_verbose 1 --mca mtl_base_verbose 1 --mca >>>> plm_base_verbose 1 --mca pls_rsh_debug 1" >>>> if [ "${WANT_CPU_AFFINITY:-NO}"= "YES"]; then >>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --report-bindings" >>>> fi >>>> fi >>>> >>>> echo "GridEngine parameters:" >>>> echo " mpirun = "`which mpirun` >>>> echo " nhosts = $NHOSTS” >>>> echo " nproc = $NSLOTS" >>>> echo " executable = $MY_EXE" >>>> echo " MPI flags = $OPENMPI_FLAGS" >>>> echo "-- begin OPENMPI run -- on " >>>> mpirun ${OPENMPI_FLAGS} python run.py >>>> echo "-- end OPENMPI run -- on “ >>>> >>>> Pls help to advise how can I solve this problem. >>>> >>>> Thanks in advance. >>>> >>>> Wei Li >>>> Ph.D student >>>> University of Delaware >>>> >>>> >>> >> >
Hi Dominik,
Thanks.
I generated the plot for Bi calculation as attached. But there are very high dense KPOINTS. I thought I change optimize it by set the parameter iterator=range as following:
result_0 = z2pack.surface.run( system=system, surface=lambda s, t: [0, s / 2, t], iterator=range(10,24,2), save_file='./results/res_0.p', load=True
But it generate same result.
Or do I need to set s limit by pos_tol?
Regards Wei Li
On Feb 16, 2017, at 5:37 PM, Dominik Gresch greschd@phys.ethz.ch wrote:
Dear Wei Li,
The surface is parametrized as a function of two parameters, as described here: http://z2pack.ethz.ch/doc/2.0/tutorial_surface.html#basic-syntax http://z2pack.ethz.ch/doc/2.0/tutorial_surface.html#basic-syntax For VASP, it is important to note that the surface must be straight in the second parameter (that is, f(t1=const., t2) describes a straight line across the full BZ, parallel to one of the reciprocal axes). This is again due to a limitation in the interface between Wannier90 and VASP, where only automatic mesh generation is possible. For your second question about the progress of the calculation: To see the progress of your calculation, it is much more helpful to look at the output of Z2Pack itself instead of the build folder or the result file. Z2Pack reports which lines it is calculating, and what the status of the convergence is.
You can improve the performance for example if you find that the iteration in a single line never converges before a specific value, by adjusting the 'iterator' parameter accordingly. The main reason it can sometimes take a long time is because of the adaptive convergence criteria. This is a tradeoff we made on purpose, because it significantly improves the reliability of the results. Of course you can turn them off, but I would strongly advise against this. All in all, performance was not so much the focus in developing Z2Pack as reliability and flexibility in using it for different purposes. Best regards,
Dominik
On 16.02.2017 22:15, Wei Li wrote:
Dear Dominik,
Thanks, I have complied wannier90 2.1 on top of VASP and I am doing the test currently.
I noticed that z2pack dynamically generate the KPOINTS files for WCC calculations. Everything runs well except that I am not sure about parameterizing the surface, which means that when the KPOINTS loop can end. So far my KPOINTS files reach to like below:
Automatic mesh 0 ! number of k-points = 0 ->automatic generation scheme Gamma ! generate a Gamma centered grid 1 1 17 ! subdivisions 0.0 0.03125 0.0 ! shift
Do I need to set a parameter to end the KPOINTS shifting, or I just need to waiting for the calculation completed as the parameters in the tutorial? So far, it realy consumes lots of computation time. Here is my results file information:
[wli9@br006 Z2]$ ls -artl results/ total 1026 -rw------- 1 wli9 mr4s90p 75853 Feb 16 16:13 res_0.p drwxr-xr-x 2 wli9 mr4s90p 33280 Feb 16 16:13 . drwxr-xr-x 6 wli9 mr4s90p 33280 Feb 16 16:13 ..
Regards Wei Li
On Feb 13, 2017, at 5:44 PM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
The Bi example should run with the supplied input files. However, you need to install a modified version of Wannier90, as explained in the installation instructions: http://z2pack.ethz.ch/doc/2.0/tutorial_installation.html#setting-up-first-pr... http://z2pack.ethz.ch/doc/2.0/tutorial_installation.html#setting-up-first-principles-codes For Wannier90 2.1 or newer this is no longer necessary because we added a specific interface to run with Z2Pack, but unfortunately VASP only supports Wannier90 version 1.x.
Best regards,
Dominik Gresch
On 13.02.2017 23:40, Wei Li wrote:
Hi Dominik,
Thanks, it runs well now.
For the given example (Bi) calculation, I need to set up the search_shells number. What’s the value you recommend? I have increased to 30 from default value 12, it still can not satisfy the B1 condition:
Unable to satisfy B1 with any of the first 30 shells Your cell might be very long, or you may have an irregular MP grid Try increasing the parameter search_shells in the win file (default=12)
Or do I need to check other parameters?
Here is my input:
num_wann = 10 num_bands = 10 spinors : true search_shells=30 num_iter 0 exclude_bands 11-15
Regards Wei Li
On Feb 11, 2017, at 4:00 PM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
If everything work correctly, Z2Pack should automatically call VASP (via the command given) and read the .mmn file after VASP has generated it. This is because Z2Pack needs to iteratively change the k-point path with which VASP is called and the overlap file is calculated.
In your case, it seems that you are submitting the VASP run to some scheduler -- this means the command returns immediately, before the .mmn file has been generated. As a result, Z2Pack tries to read the file, but it does not exist yet. You can solve this problem by submitting instead the Python process running Z2Pack, and then running VASP within that same job.
Best regards,
Dominik
On 11.02.2017 20:40, Wei Li wrote:
Dear Dominik,
From what I understand, after calling the first principle code (in my case, it is VASP), the next step is to read the .mmn file in the top level of the build file. So I give the mmn path like below:
system = z2pack.fp.System( input_files=["input/CHGCAR", "input/INCAR", "input/POSCAR", "input/POTCAR", "input/wannier90.win" ], kpt_fct=z2pack.fp.kpoint.vasp, kpt_path="KPOINTS", command="qsub /home/work/mtg/users/verali/Bi_Z2/Z2/paral.qs" mmn_path="wannier90.mmn" )
But it does not work.
Could you advise how to read the mmn file?
Regards Wei Li
> On Feb 10, 2017, at 2:59 PM, Wei Li <verali@udel.edu mailto:verali@udel.edu> wrote: > > Dear Dominik, > > Thanks for your reply. I solved this problem. > > I generated the build file as following: > > CHG CONTCAR EIGENVAL INCAR OSZICAR paral.qs.o889660 POSCAR REPORT wannier90.eig wannier90.win WAVECAR > CHGCAR DOSCAR IBZKPT KPOINTS OUTCAR PCDAT POTCAR vasprun.xml wannier90.mmn wannier90.wout XDATCAR > > There is no error. > > which script do I need to read the .mmn file to plot the result? So far, the result file is empty. > > Regards > Vera > > >> On Feb 8, 2017, at 11:16 AM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote: >> >> Dear Wei Li, >> >> It seems you are running multiple instances of the Python interpreter, due to the 'mpirun' command. There seems to be a bit of a race condition when they all try to create the 'results' folder. >> >> In general, you shouldn't run Python itself with mpirun, only the first-principles command (command="mpirun YOUR_VASP_COMMAND" in the z2pack.fp.System options). >> Best regards, >> >> Dominik Gresch >> >> On 08.02.2017 17:08, Wei Li wrote: >>> Dear Dominik Gresch, >>> >>> Thanks a lot. >>> >>> I activated py35 and it works. >>> >>> Do I need to change any parameters if I want to run the example given here: >>> https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp >>> >>> I run it directly and I got the attached output, I did not find any error in output file or log file, but the process is terminated. >>> >>> >>> Regards >>> Wei Li >>> >>> >>>> On Feb 8, 2017, at 4:53 AM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote: >>>> >>>> Dear Wei Li, >>>> >>>> The python version on which you installed Z2Pack (3.5.2) is probably not the same as you are using to run the script. Since it seems you are using a virtualenv (py35) when running Python from the command line, you should make sure the same virtualenv is also active in your job (i.e., "source py35/bin/activate" in your script). >>>> >>>> Best regards, >>>> >>>> Dominik Gresch >>>> On 08.02.2017 10:31, Wei Li wrote: >>>>> Dear Z2pack team, >>>>> >>>>> I have complied Wannier90 1.2 on top of VASP, and then I downloaded and installed z2pack like below: >>>>> >>>>> (py35) [verali@farber input]$ python >>>>> Python 3.5.2 |Anaconda 4.2.0 (64-bit)| (default, Jul 2 2016, 17:53:06) >>>>> [GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux >>>>> Type "help", "copyright", "credits" or "license" for more information. >>>>> >>> import z2pack >>>>> >>> z2pack.__version__ >>>>> '2.0.3' >>>>> >>> >>>>> >>>>> I think this means I have successfully installed z2pack, right? >>>>> >>>>> But when I run run.py script, it had the error: >>>>> >>>>> File "run.py", line 14, in <module> >>>>> Traceback (most recent call last): >>>>> File "run.py", line 14, in <module> >>>>> import z2pack >>>>> ImportError: No module named z2pack >>>>> Traceback (most recent call last): >>>>> File "run.py", line 14, in <module> >>>>> >>>>> >>>>> Below is my job file: >>>>> >>>>> >>>>> >>>>> vpkg_require intel/2015.3.187 >>>>> vpkg_require openmpi/1.8.5 >>>>> >>>>> #$ -pe mpi 40 >>>>> #$ -l m_mem_free=3G >>>>> #$ -l h_cpu=09:00:00 >>>>> ##$ -l exclusive=1 >>>>> >>>>> MY_EXE="/home/work/mtg/users/bin/vasp_ncl" >>>>> >>>>> MY_EXE_ARGS=() >>>>> >>>>> # >>>>> # Uncomment to enable lots of additional information as OpenMPI executes >>>>> # your job: >>>>> # >>>>> #SHOW_MPI_DEBUGGING=YES >>>>> >>>>> ## >>>>> ## You should NOT need to change anything after this comment. >>>>> ## >>>>> OPENMPI_FLAGS="--display-map --mca btl ^openib --mca oob_tcp_if_exclude lo,ib0 --mca btl_tcp_if_exclude lo,ib0" >>>>> if [ "${WANT_CPU_AFFINITY:-NO}" = "YES" ]; then >>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --bind-to core" >>>>> fi >>>>> if [ "${WANT_NPROC:-0}" -gt 0 ]; then >>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --np ${WANT_NPROC} --map-by node" >>>>> fi >>>>> if [ "${SHOW_MPI_DEBUGGING:-NO}" = "YES" ]; then >>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --debug-devel --debug-daemons --display-devel-map --display-devel-allocation --mca mca_verbose 1 --mca coll_base_verbose 1 --mca ras_base_verbose 1 --mca ras_gridengine_debug 1 --mca ras_gridengine_verbose 1 --mca btl_base_verbose 1 --mca mtl_base_verbose 1 --mca plm_base_verbose 1 --mca pls_rsh_debug 1" >>>>> if [ "${WANT_CPU_AFFINITY:-NO}" = "YES" ]; then >>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --report-bindings" >>>>> fi >>>>> fi >>>>> >>>>> echo "GridEngine parameters:" >>>>> echo " mpirun = "`which mpirun` >>>>> echo " nhosts = $NHOSTS” >>>>> echo " nproc = $NSLOTS" >>>>> echo " executable = $MY_EXE" >>>>> echo " MPI flags = $OPENMPI_FLAGS" >>>>> echo "-- begin OPENMPI run -- on " >>>>> mpirun ${OPENMPI_FLAGS} python run.py >>>>> echo "-- end OPENMPI run -- on “ >>>>> >>>>> Pls help to advise how can I solve this problem. >>>>> >>>>> Thanks in advance. >>>>> >>>>> Wei Li >>>>> Ph.D student >>>>> University of Delaware >>>>> >>>>> >>>> >>> >> >
Hi Wei Li,
There is clearly something wrong with these results -- there are far too many Wannier charge centers. This has nothing to do with the input parameters to Z2Pack. Instead, the input files for VASP or Wannier90 are generating too many bands. The reason for this might be that VASP automatically adds more bands when the number of bands doesn't fit the number of cores on which the calculation is running. Since Wannier90 has no way of automatically ignoring these, they are included in the calculation.
There are two ways of solving this:
1. Change the number of cores on which you run, such that only the 15 NBANDS are actually calculated
2. Check the number of bands that is actually used (you should find this in the VASP output file), and change the upper limit of "exclude_bands" in wannier90.win
Best regards, Dominik
On 20.02.2017 17:50, Wei Li wrote:
Hi Dominik,
Thanks.
I generated the plot for Bi calculation as attached. But there are very high dense KPOINTS. I thought I change optimize it by set the parameter iterator=range as following:
result_0 = z2pack.surface.run( system=system, surface=lambdas, t: [0, s / 2, t], iterator=range(10,24,2), save_file='./results/res_0.p', load=True
But it generate same result.
Or do I need to set s limit by pos_tol?
Regards Wei Li
On Feb 16, 2017, at 5:37 PM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
The surface is parametrized as a function of two parameters, as described here: http://z2pack.ethz.ch/doc/2.0/tutorial_surface.html#basic-syntax
For VASP, it is important to note that the surface must be straight in the second parameter (that is, f(t1=const., t2) describes a straight line across the full BZ, parallel to one of the reciprocal axes). This is again due to a limitation in the interface between Wannier90 and VASP, where only automatic mesh generation is possible.
For your second question about the progress of the calculation: To see the progress of your calculation, it is much more helpful to look at the output of Z2Pack itself instead of the build folder or the result file. Z2Pack reports which lines it is calculating, and what the status of the convergence is.
You can improve the performance for example if you find that the iteration in a single line never converges before a specific value, by adjusting the 'iterator' parameter accordingly. The main reason it can sometimes take a long time is because of the adaptive convergence criteria. This is a tradeoff we made on purpose, because it significantly improves the reliability of the results. Of course you can turn them off, but I would strongly advise against this. All in all, performance was not so much the focus in developing Z2Pack as reliability and flexibility in using it for different purposes.
Best regards,
Dominik
On 16.02.2017 22:15, Wei Li wrote:
Dear Dominik,
Thanks, I have complied wannier90 2.1 on top of VASP and I am doing the test currently.
I noticed that z2pack dynamically generate the KPOINTS files for WCC calculations. Everything runs well except that I am not sure about parameterizing the surface, which means that when the KPOINTS loop can end. So far my KPOINTS files reach to like below:
Automatic mesh 0 ! number of k-points = 0 ->automatic generation scheme Gamma ! generate a Gamma centered grid 1 1 17 ! subdivisions 0.0 0.03125 0.0! shift
Do I need to set a parameter to end the KPOINTS shifting, or I just need to waiting for the calculation completed as the parameters in the tutorial? So far, it realy consumes lots of computation time. Here is my results file information:
[wli9@br006 Z2]$ ls -artl results/ total 1026 -rw------- 1 wli9 mr4s90p 75853 Feb 16 16:13 res_0.p drwxr-xr-x 2 wli9 mr4s90p 33280 Feb 16 16:13 . drwxr-xr-x 6 wli9 mr4s90p 33280 Feb 16 16:13 ..
Regards Wei Li
On Feb 13, 2017, at 5:44 PM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
The Bi example should run with the supplied input files. However, you need to install a modified version of Wannier90, as explained in the installation instructions: http://z2pack.ethz.ch/doc/2.0/tutorial_installation.html#setting-up-first-pr...
For Wannier90 2.1 or newer this is no longer necessary because we added a specific interface to run with Z2Pack, but unfortunately VASP only supports Wannier90 version 1.x.
Best regards,
Dominik Gresch
On 13.02.2017 23:40, Wei Li wrote:
Hi Dominik,
Thanks, it runs well now.
For the given example (Bi) calculation, I need to set up the search_shells number. What’s the value you recommend? I have increased to 30 from default value 12, it still can not satisfy the B1 condition:
Unable to satisfy B1 with any of the first 30 shells Your cell might be very long, or you may have an irregular MP grid Try increasing the parameter search_shells in the win file (default=12)
Or do I need to check other parameters?
Here is my input:
num_wann = 10 num_bands = 10 spinors : true search_shells=30 num_iter 0 exclude_bands 11-15
Regards Wei Li
On Feb 11, 2017, at 4:00 PM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
If everything work correctly, Z2Pack should automatically call VASP (via the command given) and read the .mmn file after VASP has generated it. This is because Z2Pack needs to iteratively change the k-point path with which VASP is called and the overlap file is calculated.
In your case, it seems that you are submitting the VASP run to some scheduler -- this means the command returns immediately, before the .mmn file has been generated. As a result, Z2Pack tries to read the file, but it does not exist yet. You can solve this problem by submitting instead the Python process running Z2Pack, and then running VASP within that same job.
Best regards,
Dominik
On 11.02.2017 20:40, Wei Li wrote: > Dear Dominik, > > From what I understand, after calling the first principle code > (in my case, it is VASP), the next step is to > read the .mmn file in the top level of the build file. So I give > the mmn path like below: > > system = z2pack.fp.System( > input_files=["input/CHGCAR", "input/INCAR", "input/POSCAR", > "input/POTCAR", "input/wannier90.win"], > kpt_fct=z2pack.fp.kpoint.vasp, > kpt_path="KPOINTS", > command="qsub /home/work/mtg/users/verali/Bi_Z2/Z2/paral.qs" > mmn_path="wannier90.mmn" > ) > > But it does not work. > > Could you advise how to read the mmn file? > > Regards > Wei Li > > >> On Feb 10, 2017, at 2:59 PM, Wei Li <verali@udel.edu >> mailto:verali@udel.edu> wrote: >> >> Dear Dominik, >> >> Thanks for your reply. I solved this problem. >> >> I generated the build file as following: >> >> CHG CONTCAR EIGENVAL INCAR OSZICAR paral.qs.o889660 POSCAR >> REPORT wannier90.eig wannier90.win WAVECAR >> CHGCAR DOSCAR IBZKPT KPOINTS OUTCAR PCDAT POTCAR >> vasprun.xml wannier90.mmn wannier90.wout XDATCAR >> >> There is no error. >> >> which script do I need to read the .mmn file to plot the >> result? So far, the result file is empty. >> >> Regards >> Vera >> >> >>> On Feb 8, 2017, at 11:16 AM, Dominik Gresch >>> <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote: >>> >>> Dear Wei Li, >>> >>> It seems you are running multiple instances of the Python >>> interpreter, due to the 'mpirun' command. There seems to be a >>> bit of a race condition when they all try to create the >>> 'results' folder. >>> >>> In general, you shouldn't run Python itself with mpirun, only >>> the first-principles command (command="mpirun >>> YOUR_VASP_COMMAND" in the z2pack.fp.System options). >>> >>> Best regards, >>> >>> Dominik Gresch >>> >>> >>> On 08.02.2017 17:08, Wei Li wrote: >>>> Dear Dominik Gresch, >>>> >>>> Thanks a lot. >>>> >>>> I activated py35 and it works. >>>> >>>> Do I need to change any parameters if I want to run the >>>> example given here: >>>> https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp >>>> >>>> I run it directly and I got the attached output, I did not >>>> find any error in output file or log file, but the process is >>>> terminated. >>>> >>>> >>>> Regards >>>> Wei Li >>>> >>>> >>>>> On Feb 8, 2017, at 4:53 AM, Dominik Gresch >>>>> <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote: >>>>> >>>>> Dear Wei Li, >>>>> >>>>> The python version on which you installed Z2Pack (3.5.2) is >>>>> probably not the same as you are using to run the script. >>>>> Since it seems you are using a virtualenv (py35) when >>>>> running Python from the command line, you should make sure >>>>> the same virtualenv is also active in your job (i.e., >>>>> "source py35/bin/activate" in your script). >>>>> >>>>> Best regards, >>>>> >>>>> Dominik Gresch >>>>> >>>>> On 08.02.2017 10:31, Wei Li wrote: >>>>>> Dear Z2pack team, >>>>>> >>>>>> I have complied Wannier90 1.2 on top of VASP, and then I >>>>>> downloaded and installed z2pack like below: >>>>>> >>>>>> (py35) [verali@farber input]$ python >>>>>> Python 3.5.2 |Anaconda 4.2.0 (64-bit)| (default, Jul 2 >>>>>> 2016, 17:53:06) >>>>>> [GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux >>>>>> Type "help", "copyright", "credits" or "license" for more >>>>>> information. >>>>>> >>> import z2pack >>>>>> >>> z2pack.__version__ >>>>>> '2.0.3' >>>>>> >>> >>>>>> >>>>>> I think this means I have successfully installed z2pack, right? >>>>>> >>>>>> But when I run run.py script, it had the error: >>>>>> >>>>>> File "run.py", line 14, in <module> >>>>>> Traceback (most recent call last): >>>>>> File "run.py", line 14, in <module> >>>>>> import z2pack >>>>>> ImportError: No module named z2pack >>>>>> Traceback (most recent call last): >>>>>> File "run.py", line 14, in <module> >>>>>> >>>>>> >>>>>> *Below is my job file:* >>>>>> >>>>>> >>>>>> >>>>>> vpkg_require intel/2015.3.187 >>>>>> vpkg_require openmpi/1.8.5 >>>>>> >>>>>> #$ -pe mpi 40 >>>>>> #$ -l m_mem_free=3G >>>>>> #$ -l h_cpu=09:00:00 >>>>>> ##$ -l exclusive=1 >>>>>> >>>>>> MY_EXE="/home/work/mtg/users/bin/vasp_ncl" >>>>>> >>>>>> MY_EXE_ARGS=() >>>>>> >>>>>> # >>>>>> # Uncomment to enable lots of additional information as >>>>>> OpenMPI executes >>>>>> # your job: >>>>>> # >>>>>> #SHOW_MPI_DEBUGGING=YES >>>>>> >>>>>> ## >>>>>> ## You should NOT need to change anything after this comment. >>>>>> ## >>>>>> OPENMPI_FLAGS="--display-map --mca btl ^openib --mca >>>>>> oob_tcp_if_exclude lo,ib0 --mca btl_tcp_if_exclude lo,ib0" >>>>>> if [ "${WANT_CPU_AFFINITY:-NO}"= "YES"]; then >>>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --bind-to core" >>>>>> fi >>>>>> if [ "${WANT_NPROC:-0}"-gt 0 ]; then >>>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --np ${WANT_NPROC} >>>>>> --map-by node" >>>>>> fi >>>>>> if [ "${SHOW_MPI_DEBUGGING:-NO}"= "YES"]; then >>>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --debug-devel >>>>>> --debug-daemons --display-devel-map >>>>>> --display-devel-allocation --mca mca_verbose 1 --mca >>>>>> coll_base_verbose 1 --mca ras_base_verbose 1 --mca >>>>>> ras_gridengine_debug 1 --mca ras_gridengine_verbose 1 --mca >>>>>> btl_base_verbose 1 --mca mtl_base_verbose 1 --mca >>>>>> plm_base_verbose 1 --mca pls_rsh_debug 1" >>>>>> if [ "${WANT_CPU_AFFINITY:-NO}"= "YES"]; then >>>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --report-bindings" >>>>>> fi >>>>>> fi >>>>>> >>>>>> echo "GridEngine parameters:" >>>>>> echo " mpirun = "`which mpirun` >>>>>> echo " nhosts = $NHOSTS” >>>>>> echo " nproc = $NSLOTS" >>>>>> echo " executable = $MY_EXE" >>>>>> echo " MPI flags = $OPENMPI_FLAGS" >>>>>> echo "-- begin OPENMPI run -- on " >>>>>> mpirun ${OPENMPI_FLAGS} python run.py >>>>>> echo "-- end OPENMPI run -- on “ >>>>>> >>>>>> Pls help to advise how can I solve this problem. >>>>>> >>>>>> Thanks in advance. >>>>>> >>>>>> Wei Li >>>>>> Ph.D student >>>>>> University of Delaware >>>>>> >>>>>> >>>>> >>>> >>> >> >
Dear Dominik,
Thanks so much for all of you help. I am comfortably run z2pack now.
I did some tests for semimetal bulk and one single layer and attached are my results.
I am not sure is you are ok to help to analyze what is the topological phases of the attached two results?
I apologise for any inconvenience caused.
Regards Wei Li
On Feb 20, 2017, at 12:00 PM, Dominik Gresch greschd@phys.ethz.ch wrote:
Hi Wei Li,
There is clearly something wrong with these results -- there are far too many Wannier charge centers. This has nothing to do with the input parameters to Z2Pack. Instead, the input files for VASP or Wannier90 are generating too many bands. The reason for this might be that VASP automatically adds more bands when the number of bands doesn't fit the number of cores on which the calculation is running. Since Wannier90 has no way of automatically ignoring these, they are included in the calculation. There are two ways of solving this:
1. Change the number of cores on which you run, such that only the 15 NBANDS are actually calculated 2. Check the number of bands that is actually used (you should find this in the VASP output file), and change the upper limit of "exclude_bands" in wannier90.winBest regards, Dominik
On 20.02.2017 17:50, Wei Li wrote:
Hi Dominik,
Thanks.
I generated the plot for Bi calculation as attached. But there are very high dense KPOINTS. I thought I change optimize it by set the parameter iterator=range as following:
result_0 = z2pack.surface.run( system=system, surface=lambda s, t: [0, s / 2, t], iterator=range(10,24,2), save_file='./results/res_0.p', load=True
But it generate same result.
Or do I need to set s limit by pos_tol?
Regards Wei Li
On Feb 16, 2017, at 5:37 PM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
The surface is parametrized as a function of two parameters, as described here: http://z2pack.ethz.ch/doc/2.0/tutorial_surface.html#basic-syntax http://z2pack.ethz.ch/doc/2.0/tutorial_surface.html#basic-syntax For VASP, it is important to note that the surface must be straight in the second parameter (that is, f(t1=const., t2) describes a straight line across the full BZ, parallel to one of the reciprocal axes). This is again due to a limitation in the interface between Wannier90 and VASP, where only automatic mesh generation is possible. For your second question about the progress of the calculation: To see the progress of your calculation, it is much more helpful to look at the output of Z2Pack itself instead of the build folder or the result file. Z2Pack reports which lines it is calculating, and what the status of the convergence is.
You can improve the performance for example if you find that the iteration in a single line never converges before a specific value, by adjusting the 'iterator' parameter accordingly. The main reason it can sometimes take a long time is because of the adaptive convergence criteria. This is a tradeoff we made on purpose, because it significantly improves the reliability of the results. Of course you can turn them off, but I would strongly advise against this. All in all, performance was not so much the focus in developing Z2Pack as reliability and flexibility in using it for different purposes. Best regards,
Dominik
On 16.02.2017 22:15, Wei Li wrote:
Dear Dominik,
Thanks, I have complied wannier90 2.1 on top of VASP and I am doing the test currently.
I noticed that z2pack dynamically generate the KPOINTS files for WCC calculations. Everything runs well except that I am not sure about parameterizing the surface, which means that when the KPOINTS loop can end. So far my KPOINTS files reach to like below:
Automatic mesh 0 ! number of k-points = 0 ->automatic generation scheme Gamma ! generate a Gamma centered grid 1 1 17 ! subdivisions 0.0 0.03125 0.0 ! shift
Do I need to set a parameter to end the KPOINTS shifting, or I just need to waiting for the calculation completed as the parameters in the tutorial? So far, it realy consumes lots of computation time. Here is my results file information:
[wli9@br006 Z2]$ ls -artl results/ total 1026 -rw------- 1 wli9 mr4s90p 75853 Feb 16 16:13 res_0.p drwxr-xr-x 2 wli9 mr4s90p 33280 Feb 16 16:13 . drwxr-xr-x 6 wli9 mr4s90p 33280 Feb 16 16:13 ..
Regards Wei Li
On Feb 13, 2017, at 5:44 PM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
The Bi example should run with the supplied input files. However, you need to install a modified version of Wannier90, as explained in the installation instructions: http://z2pack.ethz.ch/doc/2.0/tutorial_installation.html#setting-up-first-pr... http://z2pack.ethz.ch/doc/2.0/tutorial_installation.html#setting-up-first-principles-codes For Wannier90 2.1 or newer this is no longer necessary because we added a specific interface to run with Z2Pack, but unfortunately VASP only supports Wannier90 version 1.x.
Best regards,
Dominik Gresch
On 13.02.2017 23:40, Wei Li wrote:
Hi Dominik,
Thanks, it runs well now.
For the given example (Bi) calculation, I need to set up the search_shells number. What’s the value you recommend? I have increased to 30 from default value 12, it still can not satisfy the B1 condition:
Unable to satisfy B1 with any of the first 30 shells Your cell might be very long, or you may have an irregular MP grid Try increasing the parameter search_shells in the win file (default=12)
Or do I need to check other parameters?
Here is my input:
num_wann = 10 num_bands = 10 spinors : true search_shells=30 num_iter 0 exclude_bands 11-15
Regards Wei Li
> On Feb 11, 2017, at 4:00 PM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote: > > Dear Wei Li, > > If everything work correctly, Z2Pack should automatically call VASP (via the command given) and read the .mmn file after VASP has generated it. This is because Z2Pack needs to iteratively change the k-point path with which VASP is called and the overlap file is calculated. > > In your case, it seems that you are submitting the VASP run to some scheduler -- this means the command returns immediately, before the .mmn file has been generated. As a result, Z2Pack tries to read the file, but it does not exist yet. You can solve this problem by submitting instead the Python process running Z2Pack, and then running VASP within that same job. > > Best regards, > > Dominik > > On 11.02.2017 20:40, Wei Li wrote: >> Dear Dominik, >> >> From what I understand, after calling the first principle code (in my case, it is VASP), the next step is to >> read the .mmn file in the top level of the build file. So I give the mmn path like below: >> >> system = z2pack.fp.System( >> input_files=["input/CHGCAR", "input/INCAR", "input/POSCAR", "input/POTCAR", "input/wannier90.win" ], >> kpt_fct=z2pack.fp.kpoint.vasp, >> kpt_path="KPOINTS", >> command="qsub /home/work/mtg/users/verali/Bi_Z2/Z2/paral.qs" >> mmn_path="wannier90.mmn" >> ) >> >> But it does not work. >> >> Could you advise how to read the mmn file? >> >> Regards >> Wei Li >> >> >>> On Feb 10, 2017, at 2:59 PM, Wei Li <verali@udel.edu mailto:verali@udel.edu> wrote: >>> >>> Dear Dominik, >>> >>> Thanks for your reply. I solved this problem. >>> >>> I generated the build file as following: >>> >>> CHG CONTCAR EIGENVAL INCAR OSZICAR paral.qs.o889660 POSCAR REPORT wannier90.eig wannier90.win WAVECAR >>> CHGCAR DOSCAR IBZKPT KPOINTS OUTCAR PCDAT POTCAR vasprun.xml wannier90.mmn wannier90.wout XDATCAR >>> >>> There is no error. >>> >>> which script do I need to read the .mmn file to plot the result? So far, the result file is empty. >>> >>> Regards >>> Vera >>> >>> >>>> On Feb 8, 2017, at 11:16 AM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote: >>>> >>>> Dear Wei Li, >>>> >>>> It seems you are running multiple instances of the Python interpreter, due to the 'mpirun' command. There seems to be a bit of a race condition when they all try to create the 'results' folder. >>>> >>>> In general, you shouldn't run Python itself with mpirun, only the first-principles command (command="mpirun YOUR_VASP_COMMAND" in the z2pack.fp.System options). >>>> Best regards, >>>> >>>> Dominik Gresch >>>> >>>> On 08.02.2017 17:08, Wei Li wrote: >>>>> Dear Dominik Gresch, >>>>> >>>>> Thanks a lot. >>>>> >>>>> I activated py35 and it works. >>>>> >>>>> Do I need to change any parameters if I want to run the example given here: >>>>> https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp >>>>> >>>>> I run it directly and I got the attached output, I did not find any error in output file or log file, but the process is terminated. >>>>> >>>>> >>>>> Regards >>>>> Wei Li >>>>> >>>>> >>>>>> On Feb 8, 2017, at 4:53 AM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote: >>>>>> >>>>>> Dear Wei Li, >>>>>> >>>>>> The python version on which you installed Z2Pack (3.5.2) is probably not the same as you are using to run the script. Since it seems you are using a virtualenv (py35) when running Python from the command line, you should make sure the same virtualenv is also active in your job (i.e., "source py35/bin/activate" in your script). >>>>>> >>>>>> Best regards, >>>>>> >>>>>> Dominik Gresch >>>>>> On 08.02.2017 10:31, Wei Li wrote: >>>>>>> Dear Z2pack team, >>>>>>> >>>>>>> I have complied Wannier90 1.2 on top of VASP, and then I downloaded and installed z2pack like below: >>>>>>> >>>>>>> (py35) [verali@farber input]$ python >>>>>>> Python 3.5.2 |Anaconda 4.2.0 (64-bit)| (default, Jul 2 2016, 17:53:06) >>>>>>> [GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux >>>>>>> Type "help", "copyright", "credits" or "license" for more information. >>>>>>> >>> import z2pack >>>>>>> >>> z2pack.__version__ >>>>>>> '2.0.3' >>>>>>> >>> >>>>>>> >>>>>>> I think this means I have successfully installed z2pack, right? >>>>>>> >>>>>>> But when I run run.py script, it had the error: >>>>>>> >>>>>>> File "run.py", line 14, in <module> >>>>>>> Traceback (most recent call last): >>>>>>> File "run.py", line 14, in <module> >>>>>>> import z2pack >>>>>>> ImportError: No module named z2pack >>>>>>> Traceback (most recent call last): >>>>>>> File "run.py", line 14, in <module> >>>>>>> >>>>>>> >>>>>>> Below is my job file: >>>>>>> >>>>>>> >>>>>>> >>>>>>> vpkg_require intel/2015.3.187 >>>>>>> vpkg_require openmpi/1.8.5 >>>>>>> >>>>>>> #$ -pe mpi 40 >>>>>>> #$ -l m_mem_free=3G >>>>>>> #$ -l h_cpu=09:00:00 >>>>>>> ##$ -l exclusive=1 >>>>>>> >>>>>>> MY_EXE="/home/work/mtg/users/bin/vasp_ncl" >>>>>>> >>>>>>> MY_EXE_ARGS=() >>>>>>> >>>>>>> # >>>>>>> # Uncomment to enable lots of additional information as OpenMPI executes >>>>>>> # your job: >>>>>>> # >>>>>>> #SHOW_MPI_DEBUGGING=YES >>>>>>> >>>>>>> ## >>>>>>> ## You should NOT need to change anything after this comment. >>>>>>> ## >>>>>>> OPENMPI_FLAGS="--display-map --mca btl ^openib --mca oob_tcp_if_exclude lo,ib0 --mca btl_tcp_if_exclude lo,ib0" >>>>>>> if [ "${WANT_CPU_AFFINITY:-NO}" = "YES" ]; then >>>>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --bind-to core" >>>>>>> fi >>>>>>> if [ "${WANT_NPROC:-0}" -gt 0 ]; then >>>>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --np ${WANT_NPROC} --map-by node" >>>>>>> fi >>>>>>> if [ "${SHOW_MPI_DEBUGGING:-NO}" = "YES" ]; then >>>>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --debug-devel --debug-daemons --display-devel-map --display-devel-allocation --mca mca_verbose 1 --mca coll_base_verbose 1 --mca ras_base_verbose 1 --mca ras_gridengine_debug 1 --mca ras_gridengine_verbose 1 --mca btl_base_verbose 1 --mca mtl_base_verbose 1 --mca plm_base_verbose 1 --mca pls_rsh_debug 1" >>>>>>> if [ "${WANT_CPU_AFFINITY:-NO}" = "YES" ]; then >>>>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --report-bindings" >>>>>>> fi >>>>>>> fi >>>>>>> >>>>>>> echo "GridEngine parameters:" >>>>>>> echo " mpirun = "`which mpirun` >>>>>>> echo " nhosts = $NHOSTS” >>>>>>> echo " nproc = $NSLOTS" >>>>>>> echo " executable = $MY_EXE" >>>>>>> echo " MPI flags = $OPENMPI_FLAGS" >>>>>>> echo "-- begin OPENMPI run -- on " >>>>>>> mpirun ${OPENMPI_FLAGS} python run.py >>>>>>> echo "-- end OPENMPI run -- on “ >>>>>>> >>>>>>> Pls help to advise how can I solve this problem. >>>>>>> >>>>>>> Thanks in advance. >>>>>>> >>>>>>> Wei Li >>>>>>> Ph.D student >>>>>>> University of Delaware >>>>>>> >>>>>>> >>>>>> >>>>> >>>> >>> >> >
Dear Wei Li,
You can analyze the topological phase of these results with the functions given in Z2Pack: http://z2pack.ethz.ch/doc/2.0/tutorial_invariants.html
This is much more reliable than trying to estimate it from the plots of the result -- indeed, the plots are usually for illustration purposes only.
Best regards,
Dominik
On 24.02.2017 21:22, Wei Li wrote:
Dear Dominik,
Thanks so much for all of you help. I am comfortably run z2pack now.
I did some tests for semimetal bulk and one single layer and attached are my results.
I am not sure is you are ok to help to analyze what is the topological phases of the attached two results?
I apologise for any inconvenience caused.
Regards Wei Li
On Feb 20, 2017, at 12:00 PM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Hi Wei Li,
There is clearly something wrong with these results -- there are far too many Wannier charge centers. This has nothing to do with the input parameters to Z2Pack. Instead, the input files for VASP or Wannier90 are generating too many bands. The reason for this might be that VASP automatically adds more bands when the number of bands doesn't fit the number of cores on which the calculation is running. Since Wannier90 has no way of automatically ignoring these, they are included in the calculation.
There are two ways of solving this:
1. Change the number of cores on which you run, such that onlythe 15 NBANDS are actually calculated
2. Check the number of bands that is actually used (you shouldfind this in the VASP output file), and change the upper limit of "exclude_bands" in wannier90.win
Best regards, Dominik
On 20.02.2017 17:50, Wei Li wrote:
Hi Dominik,
Thanks.
I generated the plot for Bi calculation as attached. But there are very high dense KPOINTS. I thought I change optimize it by set the parameter iterator=range as following:
result_0 = z2pack.surface.run( system=system, surface=lambdas, t: [0, s / 2, t], iterator=range(10,24,2), save_file='./results/res_0.p', load=True
But it generate same result.
Or do I need to set s limit by pos_tol?
Regards Wei Li
On Feb 16, 2017, at 5:37 PM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
The surface is parametrized as a function of two parameters, as described here: http://z2pack.ethz.ch/doc/2.0/tutorial_surface.html#basic-syntax
For VASP, it is important to note that the surface must be straight in the second parameter (that is, f(t1=const., t2) describes a straight line across the full BZ, parallel to one of the reciprocal axes). This is again due to a limitation in the interface between Wannier90 and VASP, where only automatic mesh generation is possible.
For your second question about the progress of the calculation: To see the progress of your calculation, it is much more helpful to look at the output of Z2Pack itself instead of the build folder or the result file. Z2Pack reports which lines it is calculating, and what the status of the convergence is.
You can improve the performance for example if you find that the iteration in a single line never converges before a specific value, by adjusting the 'iterator' parameter accordingly. The main reason it can sometimes take a long time is because of the adaptive convergence criteria. This is a tradeoff we made on purpose, because it significantly improves the reliability of the results. Of course you can turn them off, but I would strongly advise against this. All in all, performance was not so much the focus in developing Z2Pack as reliability and flexibility in using it for different purposes.
Best regards,
Dominik
On 16.02.2017 22:15, Wei Li wrote:
Dear Dominik,
Thanks, I have complied wannier90 2.1 on top of VASP and I am doing the test currently.
I noticed that z2pack dynamically generate the KPOINTS files for WCC calculations. Everything runs well except that I am not sure about parameterizing the surface, which means that when the KPOINTS loop can end. So far my KPOINTS files reach to like below:
Automatic mesh 0 ! number of k-points = 0 ->automatic generation scheme Gamma ! generate a Gamma centered grid 1 1 17 ! subdivisions 0.0 0.03125 0.0! shift
Do I need to set a parameter to end the KPOINTS shifting, or I just need to waiting for the calculation completed as the parameters in the tutorial? So far, it realy consumes lots of computation time. Here is my results file information:
[wli9@br006 Z2]$ ls -artl results/ total 1026 -rw------- 1 wli9 mr4s90p 75853 Feb 16 16:13 res_0.p drwxr-xr-x 2 wli9 mr4s90p 33280 Feb 16 16:13 . drwxr-xr-x 6 wli9 mr4s90p 33280 Feb 16 16:13 ..
Regards Wei Li
On Feb 13, 2017, at 5:44 PM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
The Bi example should run with the supplied input files. However, you need to install a modified version of Wannier90, as explained in the installation instructions: http://z2pack.ethz.ch/doc/2.0/tutorial_installation.html#setting-up-first-pr...
For Wannier90 2.1 or newer this is no longer necessary because we added a specific interface to run with Z2Pack, but unfortunately VASP only supports Wannier90 version 1.x.
Best regards,
Dominik Gresch
On 13.02.2017 23:40, Wei Li wrote: > Hi Dominik, > > Thanks, it runs well now. > > For the given example (Bi) calculation, I need to set up the > search_shells number. What’s the value you recommend? > I have increased to 30 from default value 12, it still can not > satisfy the B1 condition: > > Unable to satisfy B1 with any of the first 30 shells > Your cell might be very long, or you may have an irregular MP grid > Try increasing the parameter search_shells in the win file > (default=12) > > Or do I need to check other parameters? > > Here is my input: > > num_wann = 10 > num_bands = 10 > spinors : true > search_shells=30 > num_iter 0 > exclude_bands 11-15 > > Regards > Wei Li > > >> On Feb 11, 2017, at 4:00 PM, Dominik Gresch >> <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote: >> >> Dear Wei Li, >> >> If everything work correctly, Z2Pack should automatically call >> VASP (via the command given) and read the .mmn file after VASP >> has generated it. This is because Z2Pack needs to iteratively >> change the k-point path with which VASP is called and the >> overlap file is calculated. >> >> In your case, it seems that you are submitting the VASP run to >> some scheduler -- this means the command returns immediately, >> before the .mmn file has been generated. As a result, Z2Pack >> tries to read the file, but it does not exist yet. You can >> solve this problem by submitting instead the Python process >> running Z2Pack, and then running VASP within that same job. >> >> Best regards, >> >> Dominik >> >> >> On 11.02.2017 20:40, Wei Li wrote: >>> Dear Dominik, >>> >>> From what I understand, after calling the first principle code >>> (in my case, it is VASP), the next step is to >>> read the .mmn file in the top level of the build file. So I >>> give the mmn path like below: >>> >>> system = z2pack.fp.System( >>> input_files=["input/CHGCAR", "input/INCAR", >>> "input/POSCAR", "input/POTCAR", "input/wannier90.win"], >>> kpt_fct=z2pack.fp.kpoint.vasp, >>> kpt_path="KPOINTS", >>> command="qsub /home/work/mtg/users/verali/Bi_Z2/Z2/paral.qs" >>> mmn_path="wannier90.mmn" >>> ) >>> >>> But it does not work. >>> >>> Could you advise how to read the mmn file? >>> >>> Regards >>> Wei Li >>> >>> >>>> On Feb 10, 2017, at 2:59 PM, Wei Li <verali@udel.edu >>>> mailto:verali@udel.edu> wrote: >>>> >>>> Dear Dominik, >>>> >>>> Thanks for your reply. I solved this problem. >>>> >>>> I generated the build file as following: >>>> >>>> CHG CONTCAR EIGENVAL INCAR OSZICAR paral.qs.o889660 POSCAR >>>> REPORT wannier90.eig wannier90.win WAVECAR >>>> CHGCAR DOSCAR IBZKPT KPOINTS OUTCAR PCDAT POTCAR >>>> vasprun.xml wannier90.mmn wannier90.wout XDATCAR >>>> >>>> There is no error. >>>> >>>> which script do I need to read the .mmn file to plot the >>>> result? So far, the result file is empty. >>>> >>>> Regards >>>> Vera >>>> >>>> >>>>> On Feb 8, 2017, at 11:16 AM, Dominik Gresch >>>>> <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote: >>>>> >>>>> Dear Wei Li, >>>>> >>>>> It seems you are running multiple instances of the Python >>>>> interpreter, due to the 'mpirun' command. There seems to be >>>>> a bit of a race condition when they all try to create the >>>>> 'results' folder. >>>>> >>>>> In general, you shouldn't run Python itself with mpirun, >>>>> only the first-principles command (command="mpirun >>>>> YOUR_VASP_COMMAND" in the z2pack.fp.System options). >>>>> >>>>> Best regards, >>>>> >>>>> Dominik Gresch >>>>> >>>>> >>>>> On 08.02.2017 17:08, Wei Li wrote: >>>>>> Dear Dominik Gresch, >>>>>> >>>>>> Thanks a lot. >>>>>> >>>>>> I activated py35 and it works. >>>>>> >>>>>> Do I need to change any parameters if I want to run the >>>>>> example given here: >>>>>> https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp >>>>>> >>>>>> I run it directly and I got the attached output, I did not >>>>>> find any error in output file or log file, but the process >>>>>> is terminated. >>>>>> >>>>>> >>>>>> Regards >>>>>> Wei Li >>>>>> >>>>>> >>>>>>> On Feb 8, 2017, at 4:53 AM, Dominik Gresch >>>>>>> <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote: >>>>>>> >>>>>>> Dear Wei Li, >>>>>>> >>>>>>> The python version on which you installed Z2Pack (3.5.2) >>>>>>> is probably not the same as you are using to run the >>>>>>> script. Since it seems you are using a virtualenv (py35) >>>>>>> when running Python from the command line, you should make >>>>>>> sure the same virtualenv is also active in your job (i.e., >>>>>>> "source py35/bin/activate" in your script). >>>>>>> >>>>>>> Best regards, >>>>>>> >>>>>>> Dominik Gresch >>>>>>> >>>>>>> On 08.02.2017 10:31, Wei Li wrote: >>>>>>>> Dear Z2pack team, >>>>>>>> >>>>>>>> I have complied Wannier90 1.2 on top of VASP, and then I >>>>>>>> downloaded and installed z2pack like below: >>>>>>>> >>>>>>>> (py35) [verali@farber input]$ python >>>>>>>> Python 3.5.2 |Anaconda 4.2.0 (64-bit)| (default, Jul 2 >>>>>>>> 2016, 17:53:06) >>>>>>>> [GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux >>>>>>>> Type "help", "copyright", "credits" or "license" for more >>>>>>>> information. >>>>>>>> >>> import z2pack >>>>>>>> >>> z2pack.__version__ >>>>>>>> '2.0.3' >>>>>>>> >>> >>>>>>>> >>>>>>>> I think this means I have successfully installed z2pack, >>>>>>>> right? >>>>>>>> >>>>>>>> But when I run run.py script, it had the error: >>>>>>>> >>>>>>>> File "run.py", line 14, in <module> >>>>>>>> Traceback (most recent call last): >>>>>>>> File "run.py", line 14, in <module> >>>>>>>> import z2pack >>>>>>>> ImportError: No module named z2pack >>>>>>>> Traceback (most recent call last): >>>>>>>> File "run.py", line 14, in <module> >>>>>>>> >>>>>>>> >>>>>>>> *Below is my job file:* >>>>>>>> >>>>>>>> >>>>>>>> >>>>>>>> vpkg_require intel/2015.3.187 >>>>>>>> vpkg_require openmpi/1.8.5 >>>>>>>> >>>>>>>> #$ -pe mpi 40 >>>>>>>> #$ -l m_mem_free=3G >>>>>>>> #$ -l h_cpu=09:00:00 >>>>>>>> ##$ -l exclusive=1 >>>>>>>> >>>>>>>> MY_EXE="/home/work/mtg/users/bin/vasp_ncl" >>>>>>>> >>>>>>>> MY_EXE_ARGS=() >>>>>>>> >>>>>>>> # >>>>>>>> # Uncomment to enable lots of additional information as >>>>>>>> OpenMPI executes >>>>>>>> # your job: >>>>>>>> # >>>>>>>> #SHOW_MPI_DEBUGGING=YES >>>>>>>> >>>>>>>> ## >>>>>>>> ## You should NOT need to change anything after this comment. >>>>>>>> ## >>>>>>>> OPENMPI_FLAGS="--display-map --mca btl ^openib --mca >>>>>>>> oob_tcp_if_exclude lo,ib0 --mca btl_tcp_if_exclude lo,ib0" >>>>>>>> if [ "${WANT_CPU_AFFINITY:-NO}"= "YES"]; then >>>>>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --bind-to core" >>>>>>>> fi >>>>>>>> if [ "${WANT_NPROC:-0}"-gt 0 ]; then >>>>>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --np ${WANT_NPROC} >>>>>>>> --map-by node" >>>>>>>> fi >>>>>>>> if [ "${SHOW_MPI_DEBUGGING:-NO}"= "YES"]; then >>>>>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --debug-devel >>>>>>>> --debug-daemons --display-devel-map >>>>>>>> --display-devel-allocation --mca mca_verbose 1 --mca >>>>>>>> coll_base_verbose 1 --mca ras_base_verbose 1 --mca >>>>>>>> ras_gridengine_debug 1 --mca ras_gridengine_verbose 1 >>>>>>>> --mca btl_base_verbose 1 --mca mtl_base_verbose 1 --mca >>>>>>>> plm_base_verbose 1 --mca pls_rsh_debug 1" >>>>>>>> if [ "${WANT_CPU_AFFINITY:-NO}"= "YES"]; then >>>>>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --report-bindings" >>>>>>>> fi >>>>>>>> fi >>>>>>>> >>>>>>>> echo "GridEngine parameters:" >>>>>>>> echo " mpirun = "`which mpirun` >>>>>>>> echo " nhosts = $NHOSTS” >>>>>>>> echo " nproc = $NSLOTS" >>>>>>>> echo " executable = $MY_EXE" >>>>>>>> echo " MPI flags = $OPENMPI_FLAGS" >>>>>>>> echo "-- begin OPENMPI run -- on " >>>>>>>> mpirun ${OPENMPI_FLAGS} python run.py >>>>>>>> echo "-- end OPENMPI run -- on “ >>>>>>>> >>>>>>>> Pls help to advise how can I solve this problem. >>>>>>>> >>>>>>>> Thanks in advance. >>>>>>>> >>>>>>>> Wei Li >>>>>>>> Ph.D student >>>>>>>> University of Delaware >>>>>>>> >>>>>>>> >>>>>>> >>>>>> >>>>> >>>> >>> >> >
Dear Dr. Gresch,
Thanks so much for your help. Now I can calculate the z2 invariant and WCCs using z2pack. And I have tested some materials, like Bi2Se3, In2Se3, Graphene, etc., and the results are matched to the reports.
Again, thanks for your help.
Regards Wei Li
On Feb 25, 2017, at 6:19 PM, Dominik Gresch greschd@phys.ethz.ch wrote:
Dear Wei Li,
You can analyze the topological phase of these results with the functions given in Z2Pack: http://z2pack.ethz.ch/doc/2.0/tutorial_invariants.html http://z2pack.ethz.ch/doc/2.0/tutorial_invariants.html This is much more reliable than trying to estimate it from the plots of the result -- indeed, the plots are usually for illustration purposes only.
Best regards,
Dominik
On 24.02.2017 21:22, Wei Li wrote:
Dear Dominik,
Thanks so much for all of you help. I am comfortably run z2pack now.
I did some tests for semimetal bulk and one single layer and attached are my results.
I am not sure is you are ok to help to analyze what is the topological phases of the attached two results?
I apologise for any inconvenience caused.
Regards Wei Li
On Feb 20, 2017, at 12:00 PM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Hi Wei Li,
There is clearly something wrong with these results -- there are far too many Wannier charge centers. This has nothing to do with the input parameters to Z2Pack. Instead, the input files for VASP or Wannier90 are generating too many bands. The reason for this might be that VASP automatically adds more bands when the number of bands doesn't fit the number of cores on which the calculation is running. Since Wannier90 has no way of automatically ignoring these, they are included in the calculation. There are two ways of solving this:
1. Change the number of cores on which you run, such that only the 15 NBANDS are actually calculated 2. Check the number of bands that is actually used (you should find this in the VASP output file), and change the upper limit of "exclude_bands" in wannier90.winBest regards, Dominik
On 20.02.2017 17:50, Wei Li wrote:
Hi Dominik,
Thanks.
I generated the plot for Bi calculation as attached. But there are very high dense KPOINTS. I thought I change optimize it by set the parameter iterator=range as following:
result_0 = z2pack.surface.run( system=system, surface=lambda s, t: [0, s / 2, t], iterator=range(10,24,2), save_file='./results/res_0.p', load=True
But it generate same result.
Or do I need to set s limit by pos_tol?
Regards Wei Li
On Feb 16, 2017, at 5:37 PM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote:
Dear Wei Li,
The surface is parametrized as a function of two parameters, as described here:http://z2pack.ethz.ch/doc/2.0/tutorial_surface.html#basic-syntax http://z2pack.ethz.ch/doc/2.0/tutorial_surface.html#basic-syntax For VASP, it is important to note that the surface must be straight in the second parameter (that is, f(t1=const., t2) describes a straight line across the full BZ, parallel to one of the reciprocal axes). This is again due to a limitation in the interface between Wannier90 and VASP, where only automatic mesh generation is possible. For your second question about the progress of the calculation: To see the progress of your calculation, it is much more helpful to look at the output of Z2Pack itself instead of the build folder or the result file. Z2Pack reports which lines it is calculating, and what the status of the convergence is.
You can improve the performance for example if you find that the iteration in a single line never converges before a specific value, by adjusting the 'iterator' parameter accordingly. The main reason it can sometimes take a long time is because of the adaptive convergence criteria. This is a tradeoff we made on purpose, because it significantly improves the reliability of the results. Of course you can turn them off, but I would strongly advise against this. All in all, performance was not so much the focus in developing Z2Pack as reliability and flexibility in using it for different purposes. Best regards,
Dominik
On 16.02.2017 22:15, Wei Li wrote:
Dear Dominik,
Thanks, I have complied wannier90 2.1 on top of VASP and I am doing the test currently.
I noticed that z2pack dynamically generate the KPOINTS files for WCC calculations. Everything runs well except that I am not sure about parameterizing the surface, which means that when the KPOINTS loop can end. So far my KPOINTS files reach to like below:
Automatic mesh 0 ! number of k-points = 0 ->automatic generation scheme Gamma ! generate a Gamma centered grid 1 1 17 ! subdivisions 0.0 0.03125 0.0 ! shift
Do I need to set a parameter to end the KPOINTS shifting, or I just need to waiting for the calculation completed as the parameters in the tutorial? So far, it realy consumes lots of computation time. Here is my results file information:
[wli9@br006 Z2]$ ls -artl results/ total 1026 -rw------- 1 wli9 mr4s90p 75853 Feb 16 16:13 res_0.p drwxr-xr-x 2 wli9 mr4s90p 33280 Feb 16 16:13 . drwxr-xr-x 6 wli9 mr4s90p 33280 Feb 16 16:13 ..
Regards Wei Li
> On Feb 13, 2017, at 5:44 PM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote: > > Dear Wei Li, > > The Bi example should run with the supplied input files. However, you need to install a modified version of Wannier90, as explained in the installation instructions: http://z2pack.ethz.ch/doc/2.0/tutorial_installation.html#setting-up-first-pr... http://z2pack.ethz.ch/doc/2.0/tutorial_installation.html#setting-up-first-principles-codes > For Wannier90 2.1 or newer this is no longer necessary because we added a specific interface to run with Z2Pack, but unfortunately VASP only supports Wannier90 version 1.x. > > Best regards, > > Dominik Gresch > > On 13.02.2017 23:40, Wei Li wrote: >> Hi Dominik, >> >> Thanks, it runs well now. >> >> For the given example (Bi) calculation, I need to set up the search_shells number. What’s the value you recommend? >> I have increased to 30 from default value 12, it still can not satisfy the B1 condition: >> >> Unable to satisfy B1 with any of the first 30 shells >> Your cell might be very long, or you may have an irregular MP grid >> Try increasing the parameter search_shells in the win file (default=12) >> >> Or do I need to check other parameters? >> >> Here is my input: >> >> num_wann = 10 >> num_bands = 10 >> spinors : true >> search_shells=30 >> num_iter 0 >> exclude_bands 11-15 >> >> Regards >> Wei Li >> >> >>> On Feb 11, 2017, at 4:00 PM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote: >>> >>> Dear Wei Li, >>> >>> If everything work correctly, Z2Pack should automatically call VASP (via the command given) and read the .mmn file after VASP has generated it. This is because Z2Pack needs to iteratively change the k-point path with which VASP is called and the overlap file is calculated. >>> >>> In your case, it seems that you are submitting the VASP run to some scheduler -- this means the command returns immediately, before the .mmn file has been generated. As a result, Z2Pack tries to read the file, but it does not exist yet. You can solve this problem by submitting instead the Python process running Z2Pack, and then running VASP within that same job. >>> >>> Best regards, >>> >>> Dominik >>> >>> On 11.02.2017 20:40, Wei Li wrote: >>>> Dear Dominik, >>>> >>>> From what I understand, after calling the first principle code (in my case, it is VASP), the next step is to >>>> read the .mmn file in the top level of the build file. So I give the mmn path like below: >>>> >>>> system = z2pack.fp.System( >>>> input_files=["input/CHGCAR", "input/INCAR", "input/POSCAR", "input/POTCAR", "input/wannier90.win" ], >>>> kpt_fct=z2pack.fp.kpoint.vasp, >>>> kpt_path="KPOINTS", >>>> command="qsub /home/work/mtg/users/verali/Bi_Z2/Z2/paral.qs" >>>> mmn_path="wannier90.mmn" >>>> ) >>>> >>>> But it does not work. >>>> >>>> Could you advise how to read the mmn file? >>>> >>>> Regards >>>> Wei Li >>>> >>>> >>>>> On Feb 10, 2017, at 2:59 PM, Wei Li <verali@udel.edu mailto:verali@udel.edu> wrote: >>>>> >>>>> Dear Dominik, >>>>> >>>>> Thanks for your reply. I solved this problem. >>>>> >>>>> I generated the build file as following: >>>>> >>>>> CHG CONTCAR EIGENVAL INCAR OSZICAR paral.qs.o889660 POSCAR REPORT wannier90.eig wannier90.win WAVECAR >>>>> CHGCAR DOSCAR IBZKPT KPOINTS OUTCAR PCDAT POTCAR vasprun.xml wannier90.mmn wannier90.wout XDATCAR >>>>> >>>>> There is no error. >>>>> >>>>> which script do I need to read the .mmn file to plot the result? So far, the result file is empty. >>>>> >>>>> Regards >>>>> Vera >>>>> >>>>> >>>>>> On Feb 8, 2017, at 11:16 AM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote: >>>>>> >>>>>> Dear Wei Li, >>>>>> >>>>>> It seems you are running multiple instances of the Python interpreter, due to the 'mpirun' command. There seems to be a bit of a race condition when they all try to create the 'results' folder. >>>>>> >>>>>> In general, you shouldn't run Python itself with mpirun, only the first-principles command (command="mpirun YOUR_VASP_COMMAND" in the z2pack.fp.System options). >>>>>> Best regards, >>>>>> >>>>>> Dominik Gresch >>>>>> >>>>>> On 08.02.2017 17:08, Wei Li wrote: >>>>>>> Dear Dominik Gresch, >>>>>>> >>>>>>> Thanks a lot. >>>>>>> >>>>>>> I activated py35 and it works. >>>>>>> >>>>>>> Do I need to change any parameters if I want to run the example given here: >>>>>>> https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp https://github.com/Z2PackDev/Z2Pack/tree/master/2.0.x/examples/fp/Bi_vasp >>>>>>> >>>>>>> I run it directly and I got the attached output, I did not find any error in output file or log file, but the process is terminated. >>>>>>> >>>>>>> >>>>>>> Regards >>>>>>> Wei Li >>>>>>> >>>>>>> >>>>>>>> On Feb 8, 2017, at 4:53 AM, Dominik Gresch <greschd@phys.ethz.ch mailto:greschd@phys.ethz.ch> wrote: >>>>>>>> >>>>>>>> Dear Wei Li, >>>>>>>> >>>>>>>> The python version on which you installed Z2Pack (3.5.2) is probably not the same as you are using to run the script. Since it seems you are using a virtualenv (py35) when running Python from the command line, you should make sure the same virtualenv is also active in your job (i.e., "source py35/bin/activate" in your script). >>>>>>>> >>>>>>>> Best regards, >>>>>>>> >>>>>>>> Dominik Gresch >>>>>>>> On 08.02.2017 10:31, Wei Li wrote: >>>>>>>>> Dear Z2pack team, >>>>>>>>> >>>>>>>>> I have complied Wannier90 1.2 on top of VASP, and then I downloaded and installed z2pack like below: >>>>>>>>> >>>>>>>>> (py35) [verali@farber input]$ python >>>>>>>>> Python 3.5.2 |Anaconda 4.2.0 (64-bit)| (default, Jul 2 2016, 17:53:06) >>>>>>>>> [GCC 4.4.7 20120313 (Red Hat 4.4.7-1)] on linux >>>>>>>>> Type "help", "copyright", "credits" or "license" for more information. >>>>>>>>> >>> import z2pack >>>>>>>>> >>> z2pack.__version__ >>>>>>>>> '2.0.3' >>>>>>>>> >>> >>>>>>>>> >>>>>>>>> I think this means I have successfully installed z2pack, right? >>>>>>>>> >>>>>>>>> But when I run run.py script, it had the error: >>>>>>>>> >>>>>>>>> File "run.py", line 14, in <module> >>>>>>>>> Traceback (most recent call last): >>>>>>>>> File "run.py", line 14, in <module> >>>>>>>>> import z2pack >>>>>>>>> ImportError: No module named z2pack >>>>>>>>> Traceback (most recent call last): >>>>>>>>> File "run.py", line 14, in <module> >>>>>>>>> >>>>>>>>> >>>>>>>>> Below is my job file: >>>>>>>>> >>>>>>>>> >>>>>>>>> >>>>>>>>> vpkg_require intel/2015.3.187 >>>>>>>>> vpkg_require openmpi/1.8.5 >>>>>>>>> >>>>>>>>> #$ -pe mpi 40 >>>>>>>>> #$ -l m_mem_free=3G >>>>>>>>> #$ -l h_cpu=09:00:00 >>>>>>>>> ##$ -l exclusive=1 >>>>>>>>> >>>>>>>>> MY_EXE="/home/work/mtg/users/bin/vasp_ncl" >>>>>>>>> >>>>>>>>> MY_EXE_ARGS=() >>>>>>>>> >>>>>>>>> # >>>>>>>>> # Uncomment to enable lots of additional information as OpenMPI executes >>>>>>>>> # your job: >>>>>>>>> # >>>>>>>>> #SHOW_MPI_DEBUGGING=YES >>>>>>>>> >>>>>>>>> ## >>>>>>>>> ## You should NOT need to change anything after this comment. >>>>>>>>> ## >>>>>>>>> OPENMPI_FLAGS="--display-map --mca btl ^openib --mca oob_tcp_if_exclude lo,ib0 --mca btl_tcp_if_exclude lo,ib0" >>>>>>>>> if [ "${WANT_CPU_AFFINITY:-NO}" = "YES" ]; then >>>>>>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --bind-to core" >>>>>>>>> fi >>>>>>>>> if [ "${WANT_NPROC:-0}" -gt 0 ]; then >>>>>>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --np ${WANT_NPROC} --map-by node" >>>>>>>>> fi >>>>>>>>> if [ "${SHOW_MPI_DEBUGGING:-NO}" = "YES" ]; then >>>>>>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --debug-devel --debug-daemons --display-devel-map --display-devel-allocation --mca mca_verbose 1 --mca coll_base_verbose 1 --mca ras_base_verbose 1 --mca ras_gridengine_debug 1 --mca ras_gridengine_verbose 1 --mca btl_base_verbose 1 --mca mtl_base_verbose 1 --mca plm_base_verbose 1 --mca pls_rsh_debug 1" >>>>>>>>> if [ "${WANT_CPU_AFFINITY:-NO}" = "YES" ]; then >>>>>>>>> OPENMPI_FLAGS="${OPENMPI_FLAGS} --report-bindings" >>>>>>>>> fi >>>>>>>>> fi >>>>>>>>> >>>>>>>>> echo "GridEngine parameters:" >>>>>>>>> echo " mpirun = "`which mpirun` >>>>>>>>> echo " nhosts = $NHOSTS” >>>>>>>>> echo " nproc = $NSLOTS" >>>>>>>>> echo " executable = $MY_EXE" >>>>>>>>> echo " MPI flags = $OPENMPI_FLAGS" >>>>>>>>> echo "-- begin OPENMPI run -- on " >>>>>>>>> mpirun ${OPENMPI_FLAGS} python run.py >>>>>>>>> echo "-- end OPENMPI run -- on “ >>>>>>>>> >>>>>>>>> Pls help to advise how can I solve this problem. >>>>>>>>> >>>>>>>>> Thanks in advance. >>>>>>>>> >>>>>>>>> Wei Li >>>>>>>>> Ph.D student >>>>>>>>> University of Delaware >>>>>>>>> >>>>>>>>> >>>>>>>> >>>>>>> >>>>>> >>>>> >>>> >>> >> >