Dear Mohammad
You could try using the cudas runtime libraries provided with the pgi
compiler instead the ones of your system. Besure to have all the paths
correctly set and that no other conflicting bynary or library appears on
the path before the right ones.
For example I used these settings ( copied from the module files
provided with compiler)
export HOME_PGI='/opt/pgi' # this depends on where you installed the
compiler though
export CC=$HOME_PGI/linux86-64-llvm/2019/bin/pgcc
export FC=$HOME_PGI/linux86-64-llvm/2019/bin/pgfortran
export F90=$HOME_PGI/linux86-64-llvm/2019/bin/pgf90
export F77=$HOME_PGI/linux86-64-llvm/2019/bin/pgf77
export PATH=$HOME_PGI/linux86-64-llvm/2019/bin:$PATH
export LD_LIBRARY_PATH=$HOME_PGI/linux86-64-llvm/2019/lib:$LD_LIBRARY_PATH
export PATH=$HOME_PGI/linux86-64-llvm/2019/mpi/openmpi-3.1.3/bin:$PATH
export
LD_LIBRARY_PATH=$HOME_PGI/linux86-64-llvm/2019/mpi/openmpi-3.1.3/lib:$LD_LIBRARY_PATH
then run pgaccelinfo to obtain the compute capilities of your card (
those that yoy have to pass in the configure script with the option
-with-cuda-cc, it it the last line of the output e.g. in my case:
PGI Default Target: -ta=tesla:cc50
means that I have to use the option --with-cuda-cc=50
then run the configure ( again in my case ... )
configure --with-cuda=$HOME_PGI/pgi/linux86-64-llvm/2019/cuda/9.2
--with-cuda-runtime=9.2 --with-cuda-cc=50 --enable-openmp
--with-scalapack=no
the 2019 community edition provides versions 10.1 10.0 and 9.2 of cuda
runtime libraries so you can chose the one which fits with your card.
hope that is works for you
kind regards - Pietro Delugas
On 10/27/19 3:08 PM, Mohammad Moaddeli wrote:
Dear all,
I am trying to install q-e-gpu-qe-gpu-6.4.1a1. The following PATHs are
added in /etc/bashrc:
#### CUDA ####
export PATH=/usr/local/cuda-10.1/bin:$PATH
export PATH=/usr/local/cuda-10.1/include:$PATH
export LD_LIBRARY_PATH=/usr/local/cuda-10.1/lib64:$LD_LIBRARY_PATH
export
LD_LIBRARY_PATH=/usr/local/cuda-10.1/extras/CUPTI/lib64:$LD_LIBRARY_PATH
#### PGI ####
PGI=/opt/pgi
export PGI
PATH=/opt/pgi/linux86-64/19.4/bin:$PATH
export PATH
PATH=/opt/pgi/linux86-64/19.4/mpi/openmpi-3.1.3/bin:$PATH
export PATH
PATH=/opt/pgi/linux86-64/19.4/mpi/openmpi-3.1.3/include:$PATH
export PATH
PATH=/opt/pgi/linux86-64/19.4/mpi/openmpi-3.1.3/lib:$PATH
export PATH
MANPATH=$MANPATH:/opt/pgi/linux86-64/19.4/man
export MANPATH
and also the Graphic Card driver is installed:
[moaddeli@localhost ~]$ nvidia-smi
Sun Oct 27 16:45:25 2019
+-----------------------------------------------------------------------------+
| NVIDIA-SMI 430.50 Driver Version: 430.50 CUDA Version:
10.1 |
|-------------------------------+----------------------+----------------------+
| GPU Name Persistence-M| Bus-Id Disp.A | Volatile
Uncorr. ECC |
| Fan Temp Perf Pwr:Usage/Cap| Memory-Usage | GPU-Util
Compute M. |
|===============================+======================+======================|
| 0 GeForce GTX 108... Off | 00000000:04:00.0 On | N/A |
| 50% 28C P8 10W / 250W | 58MiB / 11175MiB | 0%
Default |
+-------------------------------+----------------------+----------------------+
+-----------------------------------------------------------------------------+
| Processes: GPU Memory |
| GPU PID Type Process name Usage |
|=============================================================================|
| 0 14528 G /usr/bin/X 39MiB |
| 0 14584 G /usr/bin/gnome-shell 16MiB |
+-----------------------------------------------------------------------------+
[root@localhost moaddeli]# lshw -numeric -C display
*-display
description: VGA compatible controller
product: GP102 [GeForce GTX 1080 Ti] [10DE:1B06]
vendor: NVIDIA Corporation [10DE]
physical id: 0
bus info: pci@0000:04:00.0
version: a1
width: 64 bits
clock: 33MHz
capabilities: pm msi pciexpress vga_controller bus_master
cap_list rom
configuration: driver=nvidia latency=0
resources: iomemory:27f0-27ef iomemory:27f0-27ef irq:89
memory:c4000000-c4ffffff memory:27fe0000000-27fefffffff
memory:27ff0000000-27ff1ffffff ioport:c000(size=128)
memory:c5000000-c507ffff
*-display
description: VGA compatible controller
product: ASPEED Graphics Family [1A03:2000]
vendor: ASPEED Technology, Inc. [1A03]
physical id: 0
bus info: pci@0000:12:00.0
version: 30
width: 32 bits
clock: 33MHz
capabilities: pm msi vga_controller cap_list
configuration: driver=ast latency=0
resources: irq:16 memory:c6000000-c6ffffff
memory:c7000000-c701ffff ioport:b000(size=128)
When I compile the code in a serial version, the executable pw.x is
created in bin directory and seems to work well:
[moaddeli@localhost test]$ /codes/qe4/q-e-gpu-qe-gpu-6.4.1a1/bin/pw.x
Program PWSCF v.6.4.1 starts on 27Oct2019 at 16:48: 2
This program is part of the open-source Quantum ESPRESSO suite
for quantum simulation of materials; please cite
"P. Giannozzi et al., J. Phys.:Condens. Matter 21 395502 (2009);
"P. Giannozzi et al., J. Phys.:Condens. Matter 29 465901 (2017);
URL http://www.quantum-espresso.org";,
in publications or presentations arising from this work. More
details at
http://www.quantum-espresso.org/quote
Serial version
Waiting for input...
however, the following error appears:
[moaddeli@localhost test]$ /codes/qe4/q-e-gpu-qe-gpu-6.4.1a1/bin/pw.x
<c.in <http://c.in> | tee c.out
0: ALLOCATE: copyin Symbol Memcpy FAILED:13(invalid device symbol)
When I compile the code in a parallel version, the executable pw.x is
created in bin directory, but it does not work:
[moaddeli@localhost ~]$ /codes/qe4/q-e-gpu-qe-gpu-6.4.1a1/bin/pw.x
[localhost.localdomain:40889] [[INVALID],INVALID] ORTE_ERROR_LOG: A
system-required executable either could not be found or was not
executable by this user in file
../../../../../orte/mca/ess/singleton/ess_singleton_module.c at line 388
[localhost.localdomain:40889] [[INVALID],INVALID] ORTE_ERROR_LOG: A
system-required executable either could not be found or was not
executable by this user in file
../../../../../orte/mca/ess/singleton/ess_singleton_module.c at line 166
--------------------------------------------------------------------------
Sorry! You were supposed to get help about:
orte_init:startup:internal-failure
But I couldn't open the help file:
/proj/pgi/linux86-64-llvm/2019/mpi/openmpi-3.1.3/share/openmpi/help-orte-runtime:
No such file or directory. Sorry!
--------------------------------------------------------------------------
--------------------------------------------------------------------------
Sorry! You were supposed to get help about:
mpi_init:startup:internal-failure
But I couldn't open the help file:
/proj/pgi/linux86-64-llvm/2019/mpi/openmpi-3.1.3/share/openmpi/help-mpi-runtime.txt:
No such file or directory. Sorry!
--------------------------------------------------------------------------
*** An error occurred in MPI_Init
*** on a NULL communicator
*** MPI_ERRORS_ARE_FATAL (processes in this communicator will now abort,
*** and potentially your MPI job)
[localhost.localdomain:40889] Local abort before MPI_INIT completed
completed successfully, but am not able to aggregate error messages,
and not able to guarantee that all other processes were killed!
Any help will be greatly appreciated.
Mohammad Moaddeli
Shiraz University
_______________________________________________
Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso)
users mailing list [email protected]
https://lists.quantum-espresso.org/mailman/listinfo/users
_______________________________________________
Quantum ESPRESSO is supported by MaX (www.max-centre.eu/quantum-espresso)
users mailing list [email protected]
https://lists.quantum-espresso.org/mailman/listinfo/users
