Re: [Pw_forum] find thread and core number
Thank you On Fri, Dec 11, 2015 at 12:40 AM, Axel Kohlmeyer wrote: > On Thu, Dec 10, 2015 at 5:25 PM, mohammed shambakey > wrote: > > Hi > > > > when running pw.x on a multi-node cluster using OpenMP+MPI, is it > possible > > to record each thread id, running on which core, in which node? > > not without using special settings and special libraries. both MPI > ranks and OpenMP thread-ids are by default 100% hardware agnostic and > are assigned to individual cores based on the kernel scheduling > algorithms, thus those can change quite rapidly. > many MPI and OpenMP implementations allow to use OS-specific APIs to > assign "affinity" of an MPI rank or a thread to specific cores (or > rather provide an assignment pattern). > > axel. > > > > > Regards > > > > -- > > Mohammed > > > > ___ > > Pw_forum mailing list > > Pw_forum@pwscf.org > > http://pwscf.org/mailman/listinfo/pw_forum > > > > -- > Dr. Axel Kohlmeyer akohl...@gmail.com http://goo.gl/1wk0 > College of Science & Technology, Temple University, Philadelphia PA, USA > International Centre for Theoretical Physics, Trieste. Italy. > ___ > Pw_forum mailing list > Pw_forum@pwscf.org > http://pwscf.org/mailman/listinfo/pw_forum > -- Mohammed ___ Pw_forum mailing list Pw_forum@pwscf.org http://pwscf.org/mailman/listinfo/pw_forum
[Pw_forum] find thread and core number
Hi when running pw.x on a multi-node cluster using OpenMP+MPI, is it possible to record each thread id, running on which core, in which node? Regards -- Mohammed ___ Pw_forum mailing list Pw_forum@pwscf.org http://pwscf.org/mailman/listinfo/pw_forum
Re: [Pw_forum] question about small benchmark and pwgui
para-speedup.sh On Thu, Nov 12, 2015 at 9:17 PM, Paolo Giannozzi wrote: > Which file exactly are you refering to? > > Paolo > > On Thu, Nov 12, 2015 at 6:52 PM, mohammed shambakey > wrote: > >> Thank you Paolo, but the file still contains "nelec" variable. >> >> Is it safe to remove it? or something else needs to be changed? >> >> Regards >> >> On Thu, Nov 12, 2015 at 10:53 AM, Paolo Giannozzi >> wrote: >> >>> >>> >>> On Tue, Nov 10, 2015 at 3:41 PM, mohammed shambakey < >>> shambak...@gmail.com> wrote: >>> >>>> >>>> 1- I'm trying to run the pwscf-small-benchmark, but it gives me an >>>> error about namelist. So, I tried to open the generated "large-test.in" >>>> file in PWgui, but it gives me the following syntax error: >>>> >>>> namelist's variable "nelec" not allowed. Skipping the rest of the file >>>> >>> >>> The updated version of the "small" benchmarks is here: >>> >>> http://qe-forge.org/gf/download/frsrelease/43/76/pwscf-small-benchmark.tar.gz >>> it runs with recent versions of QE, does not contain obsolete variable >>> "nelec" >>> >>> Paolo >>> -- >>> Paolo Giannozzi, Dept. Chemistry&Physics&Environment, >>> Univ. Udine, via delle Scienze 208, 33100 Udine, Italy >>> Phone +39-0432-558216, fax +39-0432-558222 >>> >>> ___ >>> Pw_forum mailing list >>> Pw_forum@pwscf.org >>> http://pwscf.org/mailman/listinfo/pw_forum >>> >> >> >> >> -- >> Mohammed >> >> ___ >> Pw_forum mailing list >> Pw_forum@pwscf.org >> http://pwscf.org/mailman/listinfo/pw_forum >> > > > > -- > Paolo Giannozzi, Dept. Chemistry&Physics&Environment, > Univ. Udine, via delle Scienze 208, 33100 Udine, Italy > Phone +39-0432-558216, fax +39-0432-558222 > > ___ > Pw_forum mailing list > Pw_forum@pwscf.org > http://pwscf.org/mailman/listinfo/pw_forum > -- Mohammed ___ Pw_forum mailing list Pw_forum@pwscf.org http://pwscf.org/mailman/listinfo/pw_forum
Re: [Pw_forum] question about small benchmark and pwgui
Thank you Paolo, but the file still contains "nelec" variable. Is it safe to remove it? or something else needs to be changed? Regards On Thu, Nov 12, 2015 at 10:53 AM, Paolo Giannozzi wrote: > > > On Tue, Nov 10, 2015 at 3:41 PM, mohammed shambakey > wrote: > >> >> 1- I'm trying to run the pwscf-small-benchmark, but it gives me an error >> about namelist. So, I tried to open the generated "large-test.in" file >> in PWgui, but it gives me the following syntax error: >> >> namelist's variable "nelec" not allowed. Skipping the rest of the file >> > > The updated version of the "small" benchmarks is here: > > http://qe-forge.org/gf/download/frsrelease/43/76/pwscf-small-benchmark.tar.gz > it runs with recent versions of QE, does not contain obsolete variable > "nelec" > > Paolo > -- > Paolo Giannozzi, Dept. Chemistry&Physics&Environment, > Univ. Udine, via delle Scienze 208, 33100 Udine, Italy > Phone +39-0432-558216, fax +39-0432-558222 > > ___ > Pw_forum mailing list > Pw_forum@pwscf.org > http://pwscf.org/mailman/listinfo/pw_forum > -- Mohammed ___ Pw_forum mailing list Pw_forum@pwscf.org http://pwscf.org/mailman/listinfo/pw_forum
[Pw_forum] question about small benchmark and pwgui
Hi I have two questions: 1- I'm trying to run the pwscf-small-benchmark, but it gives me an error about namelist. So, I tried to open the generated "large-test.in" file in PWgui, but it gives me the following syntax error: namelist's variable "nelec" not allowed. Skipping the rest of the file 2- I can successfully run pwgui on a local UBUNTU machine. But when I try to run it remotely on a CENTOS 6.7 machine (I logged in with ssh -X), it gives me the following error: ./pwgui.tcl: line 28: syntax error near unexpected token `(' ./pwgui.tcl: line 28: `if { [info exists env(PWGUI)] } {' Knowing that I have set PWGUI environment variable. Please help Regards -- Mohammed ___ Pw_forum mailing list Pw_forum@pwscf.org http://pwscf.org/mailman/listinfo/pw_forum
[Pw_forum] question about small size benchmarks
Hi I have a question about the small size benchmarks (pwscf-small-benchmark.tar.gz) available at ( http://qe-forge.org/gf/project/q-e/frs/?action=FrsReleaseBrowse&frs_package_id=36). My experience is in HPC, so please forgive my little knowledge about QE. Tests calculate CPU time for ELECTRONS. Why ELECTRONS, and not the PWSCF time at the end of the file? Should I use ELECTRONS time in any further performance measurements? Regards -- Mohammed ___ Pw_forum mailing list Pw_forum@pwscf.org http://pwscf.org/mailman/listinfo/pw_forum
Re: [Pw_forum] difference between cpu time and wall time
Ok, thanks. I have a test that uses 96 physical cores WITHOUT thread. The output recorded time is: PWSCF: 0h57m CPU1h44m WALL By running the same experiement with 96 physical cores and 2 threads per core (total 192 cores). The output recorded time is: PWSCF: 1h41m CPU1h39m WALL CPU time in case of 2 threads is larger than NO threads, but wall time in case of 2 threads is less than NO threads. So, I just wonder about the meaning of each criteria? and which one I should consider for performance comparison? Regards On Wed, Sep 30, 2015 at 12:42 PM, Paolo Giannozzi wrote: > Yes: "wall time" = "what the clock on the wall shows" > > Paolo > > On Wed, Sep 30, 2015 at 12:29 PM, mohammed shambakey > wrote: > >> Hi >> >> In output files, there is cpu time and wall time. Does wall time include >> cpu time plus any time for transer, I/O, and any thing else? >> >> Regards >> >> -- >> Mohammed >> >> ___ >> Pw_forum mailing list >> Pw_forum@pwscf.org >> http://pwscf.org/mailman/listinfo/pw_forum >> > > > > -- > Paolo Giannozzi, Dept. Chemistry&Physics&Environment, > Univ. Udine, via delle Scienze 208, 33100 Udine, Italy > Phone +39-0432-558216, fax +39-0432-558222 > > ___ > Pw_forum mailing list > Pw_forum@pwscf.org > http://pwscf.org/mailman/listinfo/pw_forum > -- Mohammed ___ Pw_forum mailing list Pw_forum@pwscf.org http://pwscf.org/mailman/listinfo/pw_forum
[Pw_forum] difference between cpu time and wall time
Hi In output files, there is cpu time and wall time. Does wall time include cpu time plus any time for transer, I/O, and any thing else? Regards -- Mohammed ___ Pw_forum mailing list Pw_forum@pwscf.org http://pwscf.org/mailman/listinfo/pw_forum
Re: [Pw_forum] Questions about parallelization parameters
Thank you. This is very helpful On Sun, Aug 16, 2015 at 3:47 PM, Paolo Giannozzi wrote: > 1. Correct > 2. There are some hints (calling them "guidelines" is an overstatement) in > the user guide. For most cases, you need only to consider parallelization > on k-points (-nk), > on bands using "task groups" (-ntg), on linear algebra (-nd). > 3. The maximum number of processors for linear algebra parallelization is > the same as for PW parallelization,* before *subdivision into "task > groups". The latter are used only when computing V*psi_i and |psi_i|^2 > (instead of parallelizing over PW using all processors, one parallelizes > over PW *and* band index i) > > Paolo > > > On Sun, Aug 16, 2015 at 1:19 PM, mohammed shambakey > wrote: > >> Hi >> >> >>1. As I understood from user manual, the “world” of processors is >>devided into number of “images”. Each “image” is devided into number of >>“pools”. Each “pool” is devided into number of “bands”. “PW” level depends >>on the least specified one of the previous four (i.e., if “band” is >>specified, then “PW” is automatically distributed over all processor >>available in each “band”. If “pool” is the least specified parameter, then >>“PW” is automatically distributed over all processors available in each >>“pool”, and so on). Prcoessors of each “PW” is devided into number of >>“tasks” for 3D FFT. “Linear-algebra group” specifies number of processors >>that should not exceed number of processors availabel for each “PW”, >> right? >>2. Are there guidelines for the best values for the different >>parameters (i.e., ni, nk, nb, ... etc), or I should test each problem >>individually? >>3. If my understanding is correct, the user manual says “nd is set to >>1 if ScaLAPACK is not compiled, it is set to the square integer smaller >>than or equal to half the number of processors of each pool”. But “nd” >>should not exceed “PW”; and “PW” can be smaller than “half the number of >>processors of each pool” using “nt”. If this case happens, will there be >>error? >> >> Regards >> >> -- >> Mohammed >> >> ___ >> Pw_forum mailing list >> Pw_forum@pwscf.org >> http://pwscf.org/mailman/listinfo/pw_forum >> > > > > -- > Paolo Giannozzi, Dept. Chemistry&Physics&Environment, > Univ. Udine, via delle Scienze 208, 33100 Udine, Italy > Phone +39-0432-558216, fax +39-0432-558222 > > ___ > Pw_forum mailing list > Pw_forum@pwscf.org > http://pwscf.org/mailman/listinfo/pw_forum > -- Mohammed ___ Pw_forum mailing list Pw_forum@pwscf.org http://pwscf.org/mailman/listinfo/pw_forum
[Pw_forum] Questions about parallelization parameters
Hi 1. As I understood from user manual, the “world” of processors is devided into number of “images”. Each “image” is devided into number of “pools”. Each “pool” is devided into number of “bands”. “PW” level depends on the least specified one of the previous four (i.e., if “band” is specified, then “PW” is automatically distributed over all processor available in each “band”. If “pool” is the least specified parameter, then “PW” is automatically distributed over all processors available in each “pool”, and so on). Prcoessors of each “PW” is devided into number of “tasks” for 3D FFT. “Linear-algebra group” specifies number of processors that should not exceed number of processors availabel for each “PW”, right? 2. Are there guidelines for the best values for the different parameters (i.e., ni, nk, nb, ... etc), or I should test each problem individually? 3. If my understanding is correct, the user manual says “nd is set to 1 if ScaLAPACK is not compiled, it is set to the square integer smaller than or equal to half the number of processors of each pool”. But “nd” should not exceed “PW”; and “PW” can be smaller than “half the number of processors of each pool” using “nt”. If this case happens, will there be error? Regards -- Mohammed ___ Pw_forum mailing list Pw_forum@pwscf.org http://pwscf.org/mailman/listinfo/pw_forum
Re: [Pw_forum] Computer cluster Parallel run:-Reg
Hi Suresh I'm new to Quantum_Espresso, but how many processes you specify when running the command (the "-procs" option with "poe", the "-n or -np" with "mpirun/mpiexec")? How many cores does your cluster support? Because I don't think the "-ni and -nk" options will distribute the processes by themselves. Regards On Tue, Aug 11, 2015 at 1:01 PM, Suresh A wrote: > Respected Sir/Friends, > I have formed a PC beowulf cluster (X86_64 old PC's > made on 2007) and installed quantum espresso-5.2.0 in it. Now the cluster > is running very slowly than the single core i5 processor, DDR3 RAM with > 1600MHZ speed. The following is my observation in single core i5 processor > and in beowulf cluster for the titanium dioxide supercell of size 2x2x1. > One more thing is, in cluster every node has 2GB DDR2 RAM with speed 667 > MHZ. for i5 processor 8GB with 1600 HZ. > > total cpu time spent upto now is408.7 secs (for i5 CPU having > 1600MHZ) > > total cpu time spent upto now isit is keepon running more than 1 hour > (for one single processor in PC cluster) > > total cpu time spent upto now is1402 secs ( for -ni 10 -nk 10 in PC > cluster) > > > > Is this due to RAM speed, size or anything else? > > Looking forward to your valuable suggestions > > With Regards, > A.Suresh, > Research Scholar, > Madurai Kamaraj University, > Madurai. > > ___ > Pw_forum mailing list > Pw_forum@pwscf.org > http://pwscf.org/mailman/listinfo/pw_forum > -- Mohammed ___ Pw_forum mailing list Pw_forum@pwscf.org http://pwscf.org/mailman/listinfo/pw_forum
[Pw_forum] (no subject)
Hi I trying to run the following command mpirun -np 8 pw.x < ./6x6-6+6H+F.scf.in > ./results/6x6-6+6H+F.scf.out but it fails due to missing file /C.pz-vbc.UPF. The file is downloaded and accessible in "pseudo" folder, but still have the same error. Please help Attached is the "CRASH" and "6x6-6+6H+F.scf.out" files -- Mohammed 6x6-6+6H+F.scf.out Description: Binary data CRASH Description: Binary data ___ Pw_forum mailing list Pw_forum@pwscf.org http://pwscf.org/mailman/listinfo/pw_forum
[Pw_forum] quantum_espresso 5.1.2 installation problem
Hi It is my first time to use quantum_espresso. I'm trying to install it on an hpc with redhat enterprise 6 and the following modules: gcc/4.8.1 blas/open64/64/1 mvapich2/open64/64/1.9 slurm/2.5.7 fftw3/openmpi/open64/64/3.3.3 blacs/openmpi/open64/64/1.1patch03 open64/4.5.2.1 mpich/ge/open64/64/3.0.4 scalapack/open64/64/1.8.0 configuration ends successfully, but "make all" gives me the following error: " module xdm_module ^ openf95-855 openf90: ERROR XDM_MODULE, File = xdm_dispersion.f90, Line = 11, Column = 8 The compiler has detected errors in module "XDM_MODULE". No module information file will be created for this module. REAL(DP), PARAMETER :: fact(0:8) = REAL((/1,1,2,6,24,120,720,5040,40320/),DP) ^ openf95-587 openf90: ERROR XDM_MODULE, File = xdm_dispersion.f90, Line = 71, Column = 38 The initialization expression must be a constant to be used with PARAMETER assignment for object "FACT". openf95: Open64 Fortran Version 4.5.2.1 (f14) Mon Jun 8, 2015 11:19:18 openf95: 940 source lines openf95: 2 Error(s), 0 Warning(s), 0 Other message(s), 0 ANSI(s) openf95: "explain openf95-message number" gives more information about each message make[2]: *** [xdm_dispersion.o] Error 2 make[1]: *** [pw] Error 1 make: *** [pw] Error 1 " Thanks -- Mohammed ___ Pw_forum mailing list Pw_forum@pwscf.org http://pwscf.org/mailman/listinfo/pw_forum