Impossible to define an appropriate parallel structure

[JavadH]

Dear Yambo users,

I recently changed to yambo-4.0.1 and I could successfully compile it on the supercomputer. However, running the examples provided in the webpage, I end up with the following error:

[05] Dynamical Dielectric Matrix
================================


[ERROR] STOP signal received while in :[05] Dynamical Dielectric Matrix

[ERROR]Impossible to define an appropriate parallel structure

Can anybody help me with this please? I attached the config.log file here.

Best regards,
Javad Hashemi

[Daniele Varsano]

Dear Javad,
can you please post the input/report file and the script you used to run the job?
Best,
Daniele

[JavadH]

Hello Daniele and tanks for your answer.

Here is the input:
================================
gw0 # [R GW] GoWo Quasiparticle energy levels
em1d # [R Xd] Dynamical Inverse Dielectric Matrix
HF_and_locXC # [R XX] Hartree-Fock Self-energy and Vxc
X_all_q_CPU= "1 24 1 1" # [PARALLEL] CPUs for each role
X_all_q_ROLEs= "q k c v" # [PARALLEL] CPUs roles (q,k,c,v)
X_all_q_nCPU_invert=0 # [PARALLEL] CPUs for matrix inversion
SE_CPU= "1 24 1" # [PARALLEL] CPUs for each role
SE_ROLEs= "q qp b" # [PARALLEL] CPUs roles (q,qp,b)
EXXRLvcs= 949 RL # [XX] Exchange RL components
Chimod= "hartree" # [X] IP/Hartree/ALDA/LRC/BSfxc
% GbndRnge
1 | 250 | # [GW] G[W] bands range
%
GDamping= 0.10000 eV # [GW] G[W] damping
dScStep= 0.10000 eV # [GW] Energy step to evalute Z factors
% BndsRnXd
1 | 250 | # [Xd] Polarization function bands
%
NGsBlkXd= 10 RL # [Xd] Response block size
% DmRngeXd
0.10000 | 0.10000 | eV # [Xd] Damping range
%
ETStpsXd= 100 # [Xd] Total Energy steps
% LongDrXd
1.000000 | 0.000000 | 0.000000 | # [Xd] [cc] Electric Field
%
GTermKind= "none" # [GW] GW terminator ("none","BG" Bruneval-Gonze,"BRS" Berger-Reining-Sottile)
DysSolver= "s" # [GW] Dyson Equation solver (`n`,`s`,`g`)
%QPkrange # [GW] QP generalized Kpoint/Band indices
1|500| 75|85|
%
%QPerange # [GW] QP generalized Kpoint/Energy indices
1|500| 0.0|-1.0|
%

==========================
The submit scrip:
#!/bin/bash
#SBATCH -J GW10
#SBATCH -p test
#SBATCH -n 24
#SBATCH -N 1
#SBATCH --constraint=hsw
#SBATCH -t 00:28:00
#SBATCH --mem=128000
#SBATCH --exclusive
#SBATCH -o out_%j.out
#SBATCH -e error_%j.err
srun ~/yambo/yambo-4.0.1-rev.89/bin/yambo -F yambo.in

======================================
And the header of the output:

[01] CPU structure, Files & I/O Directories
===========================================

* CPU-Threads :24(CPU)-1(threads)-1(threads@X)-1(threads@DIP)-1(threads@SE)-1(threads@RT)-1(threads@K)
* CPU-Threads :X_all_q(environment)-1 24 1 1(CPUs)-q k c v(ROLEs)
* CPU-Threads :SE(environment)-1 24 1(CPUs)-q qp b(ROLEs)
* MPI CHAINS : 4
* MPI CPU : 24
* THREADS (max): 1
* THREADS TOT(max): 24
* I/O NODES : 1
* Fragmented I/O :yes

CORE databases in .
Additional I/O in .
Communications in .
Input file is yambo.in
Report file is ./r-log_em1d_HF_and_locXC_gw0
Job string(main): log
Log files in ./LOG

[RD./SAVE//ns.db1]------------------------------------------
Bands : 250
K-points : 500
G-vectors [RL space]: 7593
Components [wavefunctions]: 6518
Symmetries [spatial+T-rev]: 2
Spinor components : 2
Spin polarizations : 1
Temperature [ev]: 0.000000
Electrons : 80.00000
WF G-vectors : 7376
Max atoms/species : 6
No. of atom species : 5
Magnetic symmetries : no
- S/N 003911 --------------------------- v.04.00.01 r.0088 -

These are the files for my own calculation. I wonder what can go wrong here.
I also would like to know about how one should choose the CPU numbers that we choose for parallelization. Should they add up to the total number of CPUs in one node, or the total CPUs that we would like to use in the calculation.

Thanks again.
Best regards,
Javad

[Daniele Varsano]

Dear Javad,
24 is not a power of two
Next, please try to attach the input and complete report file together with the error you have in the log files. You can upload them as .tar.gz file.

also would like to know about how one should choose the CPU numbers that we choose for parallelization.

There is a tutorial on the parallel structure in the web page. It is not surely complete, but anyway you can have an idea.

Should they add up to the total number of CPUs in one node, or the total CPUs that we would like to use in the calculation.

The should correspond to the amount of CPU you aim to use. Please note that they should not *add up* but the product of them ie:
X_all_q_CPU= "2 2 2 2" # [PARALLEL] CPUs for each role
X_all_q_ROLEs= "q k c v" # [PARALLEL] CPUs roles (q,k,c,v)
this correspond on a 16 MPI run.

Best,

Daniele

[JavadH]

Hi Daniele,
Thank you for your answer. I actually read the tutorial but I apparently missed it if there was mention of "power of two" there

Kind regards,
Javad