 Re: Installation error openmx 3.9 ( No.1 ) |
- Date: 2021/09/03 23:49
- Name: Naoya Yamaguchi
- Hi,
The values for `CC`, `FC` and `LIB` are for Ver. 3.8.
CC = mpicc -O3 -fopenmp -I/usr/bin -I/home/msrc/OPEN/fftw-3.3.8/api
FC = mpif90 -O3 -fopenmp -I/usr/bin -I/home/msrc/OPEN/fftw-3.3.8/api
LIB= -L/usr/lib/x86_64-linux-gnu -lfftw3 -llapack -lblas -lmpi -lmpi_mpifh -lgfortran
For Ver. 3.9, you need to add some libraries. So, there are errors of `undefined reference to`.
Please add `-lscalapack-openmpi to `LIB` as follows.
http://www.openmx-square.org/forum/patio.cgi?mode=view&no=2692
If you don't have libraries, you can install the following command.
$ sudo apt install libscalapack-openmpi-dev
(https://packages.ubuntu.com/en/bionic/libscalapack-openmpi-dev)
The above way is the most simple in your case.
Or, you can refer to the following material if you use intel softwares.
http://www.openmx-square.org/Installation_OpenMX_WSL_EN.pdf
And, please skip procedure 1-2, and follow
3. Installation of Intel one API Base Toolkit and HPC Toolkit
4. Installation of OpenMX Ver. 3.9
Dr. Yung Ting Lee also provide installation tips for Ubuntu 18.04:
https://sites.google.com/site/ytl821/home/codes
Regards,
Naoya Yamaguchi
 |
| Re: Installation error openmx 3.9 ( No.2 ) |
- Date: 2021/09/05 16:03
- Name: Vipin Kumar
- Dear Naoya,
Thank you very much for your suggestions. After installing said libraries, I successfully installed it with the command "make install". Now for applying patch3.9.6, I did the following:
1. Ran the command tar zxvf patch3.9.6.tar.gz in the 'source' directory, after copying it into 'source' directory
2. mv kpoint.in ../work/ (as given in instruction for applying patch)
3. make clean
4. make all
It shows following error:
mpicc -O3 -xHOST -ip -no-prec-div -qopenmp -I/opt/intel/mkl/include/fftw -I./elpa-2018.05.001 -c openmx.c
gcc: error: unrecognized command line option ‘-ip’; did you mean ‘-p’?
gcc: error: unrecognized command line option ‘-no-prec-div’; did you mean ‘-fno-tree-dce’?
gcc: error: unrecognized command line option ‘-qopenmp’; did you mean ‘-fopenmp’?
makefile:227: recipe for target 'openmx.o' failed
make: *** [openmx.o] Error 1
Any further help to overcome this issue. I Would be really grateful for your help.
Regards,
Vipin
|
| Re: Installation error openmx 3.9 ( No.3 ) |
- Date: 2021/09/05 16:29
- Name: Naoya Yamaguchi
- Dear Vipin,
The patch includes `makefile`, so your set parameters for `CC`, `FC` and `LIB` were overwritten.
Can you modify `makefile` again?
Regards,
Naoya Yamaguchi
|
| Re: Installation error openmx 3.9 ( No.4 ) |
- Date: 2021/09/06 13:59
- Name: Vipin <kumar.vipin118@gmail.com>
- Dear Naoya,
Thank you very much for helping hands. After modifying 'makefile', I ran 'make all'. It is showing following error:
mpicc -O3 -fopenmp -I/usr/bin -I/home/msrc/OPEN/fftw-3.3.8/api -I./elpa-2018.05.001 MulPOnly.o read_scfout.o Inputtools_kSpin.o EigenValue_Problem.o Eigen_HH.o Tools_BandCalc.o GetOrbital.o SigmaEK.o -L/usr/lib/x86_64-linux-gnu -lfftw3 -llapack -lblas -lmpi -lmpi_mpifh -lgfortran -lscalapack-openmpi -o MulPOnly
/usr/bin/ld: Tools_BandCalc.o: undefined reference to symbol 'acos@@GLIBC_2.2.5'
//lib/x86_64-linux-gnu/libm.so.6: error adding symbols: DSO missing from command line
collect2: error: ld returned 1 exit status
makefile:1154: recipe for target 'MulPOnly' failed
make: *** [MulPOnly] Error 1
It is suggesting to modify 'MulPOnly' as also given in
http://www.openmx-square.org/forum/patio.cgi?mode=view&no=2692
I can find two files: MulPOnly.c and MulPOnly.o
Do I need to edit "MulPOnly.o". Moreover, is there any particular place where I need to insert the suggestions given in
"http://www.openmx-square.org/forum/patio.cgi?mode=view&no=2692".
Because, in "MulPOnly.O" I can see only code words like @ etc. In "MulPOnly.c", program is written.
I have very less knowledge in programming. Your help would be appreciated.
Regards,
Vipin
|
| Re: Installation error openmx 3.9 ( No.5 ) |
- Date: 2021/09/06 20:14
- Name: Naoya Yamaguchi
- Dear Vipin,
The error occurs for an older version of `makefile`.
Can you use `makefile` contained in the latest patch with your modification?
Instead of it, you may solve it with a solution showed in No. 3 of the above thread.
Regards,
Naoya Yamaguchi
|
| Re: Installation error openmx 3.9 ( No.6 ) |
- Date: 2021/09/11 14:32
- Name: Vipin Kumar
- Dear Naoya,
I used the edited makefile contained in the latest patch. When I ran the command 'make all', it is showing the following error:
MulPOnly.c:312:7: warning: ignoring return value of ‘fscanf’, declared with attribute warn_unused_result [-Wunused-result]
fscanf(fp,"%d", &Data_size);
^~~~~~~~~~~~~~~~~~~~~~~~~~~
MulPOnly.c:323:2: warning: ignoring return value of ‘fscanf’, declared with attribute warn_unused_result [-Wunused-result]
fscanf(fp,"%lf",&k1); fscanf(fp,"%lf",&k2); fscanf(fp,"%lf",&k3);
^~~~~~~~~~~~~~~~~~~~
MulPOnly.c:323:31: warning: ignoring return value of ‘fscanf’, declared with attribute warn_unused_result [-Wunused-result]
fscanf(fp,"%lf",&k1); fscanf(fp,"%lf",&k2); fscanf(fp,"%lf",&k3);
^~~~~~~~~~~~~~~~~~~~
MulPOnly.c:323:60: warning: ignoring return value of ‘fscanf’, declared with attribute warn_unused_result [-Wunused-result]
fscanf(fp,"%lf",&k1); fscanf(fp,"%lf",&k2); fscanf(fp,"%lf",&k3);
^~~~~~~~~~~~~~~~~~~~
MulPOnly.c:327:2: warning: ignoring return value of ‘fscanf’, declared with attribute warn_unused_result [-Wunused-result]
fscanf(fp,"%d",&NBand_Cc[j]);
^~~~~~~~~~~~~~~~~~~~~~~~~~~~
mpicc -O3 -fopenmp -I/usr/bin -I/home/msrc/OPEN/fftw-3.3.8/api -I./elpa-2018.05.001 MulPOnly.o read_scfout.o Inputtools_kSpin.o EigenValue_Problem.o Eigen_HH.o Tools_BandCalc.o GetOrbital.o SigmaEK.o -L/usr/lib/x86_64-linux-gnu -lfftw3 -llapack -lblas -lmpi -lmpi_mpifh -lgfortran -lscalapack-openmpi -o MulPOnly
/usr/bin/ld: Tools_BandCalc.o: undefined reference to symbol 'acos@@GLIBC_2.2.5'
//lib/x86_64-linux-gnu/libm.so.6: error adding symbols: DSO missing from command line
collect2: error: ld returned 1 exit status
makefile:1159: recipe for target 'MulPOnly' failed
make: *** [MulPOnly] Error 1
Could you please help me in solving this issue?
Regards,
Vipin
|
| Re: Installation error openmx 3.9 ( No.7 ) |
- Date: 2021/09/12 01:35
- Name: Naoya Yamaguchi
- Dear Vipin,
Thank you for your careful checking.
I've found that the correction of http://www.openmx-square.org/forum/patio.cgi?mode=view&no=2692 is not included.
Can you add `-lm` to `LIB?
You can also solve it by modifying `makefile` as shown in the above thread.
Regards,
Naoya Yamaguchi
|
| Re: Installation error openmx 3.9 ( No.8 ) |
- Date: 2021/09/13 20:05
- Name: Vipin Kumar
- Dear Naoya,
Thank you very much for help. After incorporating your suggestion I did the following:
1. make clean
2. make all: it completed without any hassle
3. make install: it simply copied openmx executable to the work directory
I ran the runtest. Following is output:
1 input_example/Benzene.dat Elapsed time(s)= 8.25 diff Utot= 0.000000000038 diff Force= 0.000000000005
2 input_example/C60.dat Elapsed time(s)= 40.48 diff Utot= 0.000000000001 diff Force= 0.000000000001
3 input_example/CO.dat Elapsed time(s)= 17.78 diff Utot= 0.000000000096 diff Force= 0.000000000283
4 input_example/Cr2.dat Elapsed time(s)= 16.96 diff Utot= 0.000000000913 diff Force= 0.000000000171
5 input_example/Crys-MnO.dat Elapsed time(s)= 93.58 diff Utot= 0.000000000001 diff Force= 0.000000000043
6 input_example/GaAs.dat Elapsed time(s)= 85.48 diff Utot= 0.000000000001 diff Force= 0.000000000000
7 input_example/Glycine.dat Elapsed time(s)= 10.16 diff Utot= 0.000000000001 diff Force= 0.000000000001
8 input_example/Graphite4.dat Elapsed time(s)= 16.42 diff Utot= 0.000000000014 diff Force= 0.000000000041
9 input_example/H2O-EF.dat Elapsed time(s)= 8.05 diff Utot= 0.000000000000 diff Force= 0.000000000001
10 input_example/H2O.dat Elapsed time(s)= 6.26 diff Utot= 0.000000000000 diff Force= 0.000000000511
11 input_example/HMn.dat Elapsed time(s)= 31.83 diff Utot= 0.000000000132 diff Force= 0.000000000001
12 input_example/Methane.dat Elapsed time(s)= 5.62 diff Utot= 0.000000000008 diff Force= 0.000000000001
13 input_example/Mol_MnO.dat Elapsed time(s)= 18.64 diff Utot= 0.000000000371 diff Force= 0.000000000011
14 input_example/Ndia2.dat Elapsed time(s)= 12.98 diff Utot= 0.000000000001 diff Force= 0.000000000000
It looks fine for me as per instruction manual.
I also checked NEGF calculation for 8zgnr given in negf-example directory. SCF for leads are as follows:
For lead L:
SCF history at MD= 1
***********************************************************
***********************************************************
SCF= 1 NormRD= 1.000000000000 Uele= -50.509598396289
SCF= 2 NormRD= 12.116445673607 Uele= -50.509598396289
SCF= 3 NormRD= 12.110280781020 Uele= -50.523504684768
SCF= 4 NormRD= 11.469376752307 Uele= -51.979659225441
SCF= 5 NormRD= 10.881951991091 Uele= -53.308119968435
SCF= 6 NormRD= 10.341659348821 Uele= -54.522029062403
SCF= 7 NormRD= 9.843350071416 Uele= -55.632585449512
SCF= 8 NormRD= 9.382574350542 Uele= -56.649846039274
SCF= 9 NormRD= 8.955401848504 Uele= -57.582715126573
SCF= 10 NormRD= 7.385653373432 Uele= -58.438999017479
SCF= 11 NormRD= 7.471385635873 Uele= -59.476476310592
SCF= 12 NormRD= 9.104097295672 Uele= -60.722415506678
SCF= 13 NormRD= 11.000725243113 Uele= -59.212913768197
SCF= 14 NormRD= 9.560202087028 Uele= -78.995131274969
SCF= 15 NormRD= 7.534719293456 Uele= -77.036238727969
SCF= 16 NormRD= 5.316132644972 Uele= -74.902029461579
SCF= 17 NormRD= 4.182179571855 Uele= -73.842521451759
SCF= 18 NormRD= 2.732399706817 Uele= -72.462319825103
Please see SCF 12 and 13. Similar situation for Lead-R.
For NEGF:
SCF history at MD= 1
***********************************************************
***********************************************************
SCF= 1 NormRD= 1.000000000000 Uele= 0.000000000000
SCF= 2 NormRD= 0.196767625943 Uele= 0.000000000000
SCF= 3 NormRD= 0.188184488568 Uele= 0.000000000000
SCF= 4 NormRD= 0.161500511971 Uele= 0.000000000000
SCF= 5 NormRD= 0.137781389912 Uele= 0.000000000000
SCF= 6 NormRD= 0.116322244486 Uele= 0.000000000000
SCF= 7 NormRD= 0.112000122982 Uele= 0.000000000000
SCF= 8 NormRD= 0.100836190241 Uele= 0.000000000000
SCF= 9 NormRD= 0.094366407256 Uele= 0.000000000000
SCF= 10 NormRD= 0.088462169566 Uele= 0.000000000000
SCF= 11 NormRD= 0.083320129157 Uele= 0.000000000000
SCF= 12 NormRD= 0.078679082740 Uele= 0.000000000000
SCF= 13 NormRD= 0.074459455697 Uele= 0.000000000000
SCF= 14 NormRD= 0.070584433617 Uele= 0.000000000000
SCF= 15 NormRD= 0.067002814241 Uele= 0.000000000000
SCF= 16 NormRD= 0.063673295293 Uele= 0.000000000000
SCF= 17 NormRD= 0.060566001149 Uele= 0.000000000000
SCF= 18 NormRD= 0.057656263653 Uele= 0.000000000000
SCF= 19 NormRD= 0.054923983825 Uele= 0.000000000000
SCF= 20 NormRD= 0.052352429033 Uele= 0.000000000000
SCF= 21 NormRD= 1.000000000000 Uele= 0.000000000000
SCF= 22 NormRD= 0.061720928024 Uele= 0.000000000000
SCF= 23 NormRD= 0.055165592920 Uele= 0.000000000000
SCF= 24 NormRD= 0.054362194644 Uele= 0.000000000000
Please see SCF 21 next step to NEGF.SCF.Iter.Band (= 20)
I am not sure about it. Could you please suggest me whether these are OK or there is some problem in installation.
Regards,
Vipin
|
| Re: Installation error openmx 3.9 ( No.9 ) |
- Date: 2021/09/14 03:14
- Name: Naoya Yamaguchi
- Dear Vipin,
You can compare your computational results with out files contained `./negf_example/`.
>Please see SCF 21 next step to NEGF.SCF.Iter.Band (= 20)
This behavior might be related to No.5 in http://www.openmx-square.org/forum/patio.cgi?mode=view&no=2875 .
Regards,
Naoya Yamaguchi
|
| Re: Installation error openmx 3.9 ( No.10 ) |
- Date: 2021/09/14 20:19
- Name: Vipin Kumar
- Dear Naoya,
Thank you very much. I ran the "mpirun -np 8 openmx -runtestNEGF -nt 2". The output is follows:
1 negf_example/Lead-Au-Chain-NC. Elapsed time(s)= 50.24 diff Utot= 0.000000002012 diff Force= 0.000000000088
diff EigenChannel= 0.000000000000 diff CurrentDensity= 0.000000000000
2 negf_example/Lead-Chain.dat Elapsed time(s)= 11.48 diff Utot= 0.000000000067 diff Force= 0.000000000001
diff EigenChannel= 0.000000000000 diff CurrentDensity= 0.000000000000
3 negf_example/Lead-Graphene.dat Elapsed time(s)= 29.78 diff Utot= 0.000000000007 diff Force= 0.000000000001
diff EigenChannel= 0.000000000000 diff CurrentDensity= 0.000000000000
4 negf_example/Lead-L-8ZGNR-NC.d Elapsed time(s)= 1150.73 diff Utot= 0.000038204265 diff Force= 0.000000255689
diff EigenChannel= 0.000000000000 diff CurrentDensity= 0.000000000000
5 negf_example/Lead-L-8ZGNR.dat Elapsed time(s)= 481.32 diff Utot= 0.000000000002 diff Force= 0.000000000007
diff EigenChannel= 0.000000000000 diff CurrentDensity= 0.000000000000
6 negf_example/Lead-R-8ZGNR-NC.d Elapsed time(s)= 1071.34 diff Utot= 0.000045529356 diff Force= 0.000000197170
diff EigenChannel= 0.000000000000 diff CurrentDensity= 0.000000000000
7 negf_example/Lead-R-8ZGNR.dat Elapsed time(s)= 482.12 diff Utot= 0.000000000065 diff Force= 0.000000000002
diff EigenChannel= 0.000000000000 diff CurrentDensity= 0.000000000000
8 negf_example/NEGF-8ZGNR-0.3.da Elapsed time(s)= 6532.98 diff Utot= 0.928337063925 diff Force= 0.027250584867
diff EigenChannel= 0.006449277040 diff CurrentDensity= 0.000003818326
Steps 4, 6 and 8 are not consistent with the reference results.
Moreover, in between calculation it shows following:
TFNAN= 2442 Average FNAN= 29.78049
TSNAN= 0 Average SNAN= 0.00000
<truncation> CpyCell= 1 ct_AN= 1 FNAN SNAN 24 0
<truncation> CpyCell= 1 ct_AN= 2 FNAN SNAN 26 0
<truncation> CpyCell= 1 ct_AN= 3 FNAN SNAN 31 0
<truncation> CpyCell= 1 ct_AN= 4 FNAN SNAN 32 0
<truncation> CpyCell= 1 ct_AN= 5 FNAN SNAN 33 0
<truncation> CpyCell= 1 ct_AN= 6 FNAN SNAN 33 0
<truncation> CpyCell= 1 ct_AN= 7 FNAN SNAN 33 0
<truncation> CpyCell= 1 ct_AN= 8 FNAN SNAN 33 0
<truncation> CpyCell= 1 ct_AN= 9 FNAN SNAN 33 0
<truncation> CpyCell= 1 ct_AN= 10 FNAN SNAN 33 0
<truncation> CpyCell= 1 ct_AN= 11 FNAN SNAN 33 0
<truncation> CpyCell= 1 ct_AN= 12 FNAN SNAN 33 0
<truncation> CpyCell= 1 ct_AN= 13 FNAN SNAN 32 0
<truncation> CpyCell= 1 ct_AN= 14 FNAN SNAN 31 0
<truncation> CpyCell= 1 ct_AN= 15 FNAN SNAN 26 0
Now for the calculation of "negf_example/NEGF-8ZGNR-NC.dat", it is giving following up to 20 SCF initial steps:
******************* MD= 1 SCF= 1 *******************
<TRAN_Set_Electrode_Grid>
<TRAN_Allocate_Lead_Region>
<TRAN_Allocate_Cregion: NUM_c=760 NUM_e=336 336>
<TRAN_Add_Density_Lead>
<TRAN_Add_ADensity_Lead>
<TRAN_Poisson_FD> Solving Poisson's equation...
<NEGF> Solving the eigenvalue problem...
KGrids1: 0.00000
KGrids2: 0.00000
KGrids3: 0.00000
<Band_DFT> Eigen, time=1.215302
<Band_DFT> DM, time=0.000000
BLACS WARNING 'Trying to free non-existent system context handle 1'
from {-1,-1}, pnum=5, Contxt=-1, on line 20 of file '/build/scalapack-5is7dY/scalapack-2.0.2/BLACS/SRC/free_handle_.c'.
BLACS WARNING 'Trying to free non-existent system context handle 1'
from {-1,-1}, pnum=6, Contxt=-1, on line 20 of file '/build/scalapack-5is7dY/scalapack-2.0.2/BLACS/SRC/free_handle_.c'.
BLACS WARNING 'Trying to free non-existent system context handle 1'
from {-1,-1}, pnum=7, Contxt=-1, on line 20 of file '/build/scalapack-5is7dY/scalapack-2.0.2/BLACS/SRC/free_handle_.c'.
BLACS WARNING 'Trying to free non-existent system context handle 1'
from {-1,-1}, pnum=4, Contxt=-1, on line 20 of file '/build/scalapack-5is7dY/scalapack-2.0.2/BLACS/SRC/free_handle_.c'.
BLACS WARNING 'Trying to free non-existent system context handle 1'
from {-1,-1}, pnum=0, Contxt=-1, on line 20 of file '/build/scalapack-5is7dY/scalapack-2.0.2/BLACS/SRC/free_handle_.c'.
BLACS WARNING 'Trying to free non-existent system context handle 1'
from {-1,-1}, pnum=2, Contxt=-1, on line 20 of file '/build/scalapack-5is7dY/scalapack-2.0.2/BLACS/SRC/free_handle_.c'.
BLACS WARNING 'Trying to free non-existent system context handle 1'
from {-1,-1}, pnum=1, Contxt=-1, on line 20 of file '/build/scalapack-5is7dY/scalapack-2.0.2/BLACS/SRC/free_handle_.c'.
BLACS WARNING 'Trying to free non-existent system context handle 1'
from {-1,-1}, pnum=3, Contxt=-1, on line 20 of file '/build/scalapack-5is7dY/scalapack-2.0.2/BLACS/SRC/free_handle_.c'.
1 C MulP 1.68 1.46 sum 3.15 diff 0.22 (170.16 270.00) Ml 0.00 (156.11 77.88) Ml+s 0.22 (170.16 -90.00)
2 C MulP 1.76 1.50 sum 3.26 diff 0.26 (174.08 -90.00) Ml 0.00 ( 20.78 -83.49) Ml+s 0.26 (174.08 -90.00)
3 C MulP 1.59 1.33 sum 2.92 diff 0.26 (174.53 -90.00) Ml 0.00 (161.07 254.48) Ml+s 0.26 (174.53 270.00)
4 C MulP 1.51 1.26 sum 2.78 diff 0.25 (178.81 -90.00) Ml 0.00 ( 52.40 262.09) Ml+s 0.25 (178.81 -90.00)
5 C MulP 1.53 1.30 sum 2.83 diff 0.24 (176.26 -90.00) Ml 0.00 ( 67.94 96.36) Ml+s 0.24 (176.26 270.00)
6 C MulP 1.59 1.32 sum 2.91 diff 0.27 (178.12 90.00) Ml 0.00 (121.05 -78.72) Ml+s 0.27 (178.12 90.00)
I don't know why this is happening. Any advice on this and help.
Regards,
Vipin
|
| Re: Installation error openmx 3.9 ( No.11 ) |
- Date: 2021/09/14 20:53
- Name: Naoya Yamaguchi
- Dear Vipin,
The warning has no effect on results.
You may avoid the warning by using another libraries about linear algebra.
For example, Can you try installing `libopenblas-base` as follows.
$ sudo apt install libopenblas-base
Or, you can use Intel MKL. You can refer to the following.
http://www.openmx-square.org/Installation_OpenMX_WSL_EN.pdf
(See "3. Installation of Intel one API Base Toolkit and HPC Toolkit".)
https://drive.google.com/file/d/11PBW0woSH4j03T80oEwCd_ObCzvhNbEd/view
If you change libraries as above, you need to build OpenMX again after `make clean`.
Regards,
Naoya Yamaguchi
|
| Re: Installation error openmx 3.9 ( No.12 ) |
- Date: 2021/09/22 13:54
- Name: Vipin <kumar.vipin118@gmail.com>
- Dear Naoya,
Thank you very much for help. So far looks fine everything after checking automatic runtestresult. However, when I am running test for NEGF it is more taking more time than usual. For example, I am running now NEGF at V=-0.3V as given in negf_example directory. It took more than 24 hours to complete SCF+NEGF convergence for accuracy of the order of 10^-7. Is there any way to accelerate it.
Regards,
Vipin
|
| Re: Installation error openmx 3.9 ( No.13 ) |
- Date: 2021/09/22 21:07
- Name: Naoya Yamaguchi
- Dear Vipin,
>Is there any way to accelerate it.
Increasing MPI processes, that is, using more CPUs can accelerate calculations.
Another way is finding the optimal setting for compilation.
Regards,
Naoya Yamaguchi
|
|
This thread is locked.Only browsing is available.
Installation error openmx 3.9