Top Page > Browsing
installation error:config-f90.h
Date: 2022/09/02 00:20
Name: mahendra

Dear OpenMX developers

I am have attempted to install openmx on Ubuntu 22.04.

My compilation is based on following configuration

MKLROOT = /opt/intel/oneapi/mkl/2022.1.0
CC = mpicc -O3 -fopenmp -I${MKLROOT}/include/fftw -I${MKLROOT}/include
FC = mpif90 -O3 -fopenmp -I${MKLROOT}/include/fftw -fallow-argument-mismatch
LIB= -L${MKLROOT}/include/fftw -lfftw3 -L$MKLROOT/lib/intel64 -lmkl_blacs_intelmpi_lp64 -lmkl_scalapack_lp64 -lmkl_intel_lp64 -lmkl_core -lmkl_intel_thread -lpthread -lifcore

-------------------------------------------------------------------------

I am facing the following error
-----------------------------------------------------------------------
46 | !> \brief Fortran module which defines the datatypes used in ELPA
| 1~~~~~~~~~~~~~
Fatal Error: config-f90.h: No such file or directory
compilation terminated.
make: *** [makefile:990: mod_precision.o] Error 1
-------------------------------------------------------------------------
elpa1.o also been an issue but after I updated the elpa folder to elpa-2022.05.001, the error disappeared.

I would appreciate if someone could suggest a solution.

Mahendra
Vellore Institute of Technology
Vellore-India
e
Page: [1]

Re: installation error:config-f90.h ( No.1 )
Date: 2022/09/02 12:42
Name: Naoya Yamaguchi

Hi,

Can you show a full build log when the error appears?

Regards,
Naoya Yamaguchi
e
Re: installation error:config-f90.h ( No.2 )
Date: 2022/09/02 18:49
Name: mahendra

Dear Naoya Yamaguchi,
--------------------------------------------------------------------------------------------------------------------------
Here is my error log
f77func.h:63:27: note: in definition of macro eF77_NAMEf
63 | #define F77_NAME(a,A) a ## _
| ^
Band_DFT_Col_Optical_ScaLAPACK.c:927:5: warning: implicit declaration of function eCblacs_barrierf [-Wimplicit-function-declaration]
927 | Cblacs_barrier(ictxt2,"A");
| ^~~~~~~~~~~~~~
In file included from openmx_common.h:120,
from Band_DFT_Col_Optical_ScaLAPACK.c:18:
Band_DFT_Col_Optical_ScaLAPACK.c:929:14: warning: implicit declaration of function epzgemm_f; did you mean ezgemm_f? [-Wimplicit-function-declaration]
929 | F77_NAME(pzgemm,PZGEMM)("N","N",&n,&n,&n,&alpha,Hs,&ONE,&ONE,descH,Ss,&ONE,&ONE,descS,&beta,Cs,&ONE,&ONE,descC);
| ^~~~~~
f77func.h:63:27: note: in definition of macro eF77_NAMEf
63 | #define F77_NAME(a,A) a ## _
| ^
In file included from Band_DFT_Col_Optical_ScaLAPACK.c:18:
Band_DFT_Col_Optical_ScaLAPACK.c: At top level:
openmx_common.h:3559:13: warning: inline function eSpherical_Besself declared but never defined
3559 | inline void Spherical_Bessel( double x, int lmax, double *sb, double *dsb ) ;
| ^~~~~~~~~~~~~~~~
NBO_Cluster.c: In function eCalc_NAO_Clusterf:
NBO_Cluster.c:5583:28: warning: argument 1 range [18446744056529682440, 18446744073709551544] exceeds maximum object size 9223372036854775807 [-Walloc-size-larger-than=]
5583 | Table_NBO = (int**)malloc(sizeof(int*)*(atomnum*10+1));
| ^~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
In file included from NBO_Cluster.c:19:
/usr/include/stdlib.h:540:14: note: in a call to allocation function emallocf declared here
540 | extern void *malloc (size_t __size) __THROW __attribute_malloc__
| ^~~~~~
In file included from openmx_common.h:120,
from Cluster_DFT_Optical_ScaLAPACK.c:18:
Cluster_DFT_Optical_ScaLAPACK.c: In function eCluster_collinear_Opticalf:
Cluster_DFT_Optical_ScaLAPACK.c:375:14: warning: implicit declaration of function esolve_evp_real_f [-Wimplicit-function-declaration]
375 | F77_NAME(solve_evp_real,SOLVE_EVP_REAL)(&n, &n, Cs, &na_rows, &ko[0][1], Ss, &na_rows, &nblk, &mpi_comm_rows_int, &mpi_comm_cols_int);
| ^~~~~~~~~~~~~~
f77func.h:63:27: note: in definition of macro eF77_NAMEf
63 | #define F77_NAME(a,A) a ## _
| ^
Cluster_DFT_Optical_ScaLAPACK.c:541:3: warning: implicit declaration of function eCblacs_barrierf [-Wimplicit-function-declaration]
541 | Cblacs_barrier(ictxt1,"A");
| ^~~~~~~~~~~~~~
In file included from openmx_common.h:120,
from Cluster_DFT_Optical_ScaLAPACK.c:18:
Cluster_DFT_Optical_ScaLAPACK.c:542:12: warning: implicit declaration of function epdgemm_f [-Wimplicit-function-declaration]
542 | F77_NAME(pdgemm,PDGEMM)("N","N",&n,&n,&n,&alpha,Hs,&ONE,&ONE,descH,Ss,&ONE,&ONE,descS,&beta,Cs,&ONE,&ONE,descC);
| ^~~~~~
f77func.h:63:27: note: in definition of macro eF77_NAMEf
63 | #define F77_NAME(a,A) a ## _
| ^
In file included from Cluster_DFT_Optical_ScaLAPACK.c:18:
Cluster_DFT_Optical_ScaLAPACK.c: At top level:
openmx_common.h:3559:13: warning: inline function eSpherical_Besself declared but never defined
3559 | inline void Spherical_Bessel( double x, int lmax, double *sb, double *dsb ) ;
| ^~~~~~~~~~~~~~~~
make: *** No rule to make target 'elpa-2022.05.001/elpa_utilities.F90', needed by 'elpa_utilities.o'. Stop.
make: *** Waiting for unfinished jobs....
./elpa-2022.05.001/mod_precision.F90:46:2:

46 | !> \brief Fortran module which defines the datatypes used in ELPA
| 1~~~~~~~~~~~~~
Fatal Error: config-f90.h: No such file or directory
compilation terminated.
make: *** [makefile:990: mod_precision.o] Error 1
--------------------------------------------------------------------------------------------------------------------------------------

mahendra
e
Re: installation error:config-f90.h ( No.3 )
Date: 2022/09/02 19:24
Name: Naoya Yamaguchi

Dear mahendra,

How about using the ELPA embedded in the source directory of the OpenMX, instead of `elpa-2022.05.001`?

Regards,
Naoya Yamaguchi
e
Re: installation error:config-f90.h ( No.4 )
Date: 2022/09/04 18:36
Name: mahendra

Dear Naoya Yamaguchi,

As you suggested I have used the elpa which came with openmx3.9.
Now elpa issue appears to be solved...
And i couldn't link the ifcore library.

The following is current error...
________________________________________________________
/usr/bin/ld: cannot find -lifcore: No such file or directory
collect2: error: ld returned 1 exit status
make: *** [makefile:211: openmx] Error 1
--------------------------------------------------------
Mahendra
e
Re: installation error:config-f90.h ( No.5 )
Date: 2022/09/04 21:00
Name: Naoya Yamaguchi

Dear Mahendra,

The `-lifcore` is necessary only when you use the Intel compiler. If you use the Intel compiler, you need to set the compiler commands to `CC` and `FC` as appropriate. As for the oneAPI, you can refer to pp.18-20 of http://www.openmx-square.org/Installation_OpenMX_WSL_EN.pdf .

Regards,
Naoya Yamaguchi
e
Re: installation error:config-f90.h ( No.6 )
Date: 2022/09/04 23:15
Name: mahendra

Dear Naoya Yamaguchi,

After installing the intel-hpckit...the installation went smoothly.
Thanks a lot.

Mahendra
e

Page: [1]

Thread Title (must) Move the thread to the top
Your Name (must)
E-Mail (must)
URL
Password (used in modification of the submitted text)
Comment (must)

   Save Cookie