Dear group!<br><br>I am experiencing the following problem with pw.x. When I run a job on a single core with 8 processors, the calculation does not exceed the available RAM, i.e. it works. When I run the same job on two cores, the calculation crashes (mpierun: Forwarding signal 12 to job) [relevant log-parts see below]. However, I can compute this job on two cores with a smaller k-point sampling, hence I suspect that it has to do somehow with the memory demands/distribution. I am using the "Davidson iterative diagonalization" as a minimizer, with just settings for the thresholds (convergence, 1st iterative diagonalization).<br>
Can you please point me into the right direction? <br>With thanks<br>Yours Wolfgang<br>---<br>Parallel version (MPI)<br><br>Number of processors in use: 16<br>R & G space division: proc/pool = 16<br>...<br>Subspace diagonalization in iterative solution of the eigenvalue problem:<br>
a parallel distributed memory algorithm will be used,<br>eigenstates matrixes will be distributed block like on<br>ortho sub-group = 4* 4 procs<br>...<br>Planes per process (thick) : nr3 = 90 npp = 6 ncplane = 8100<br><br>
Proc/ planes cols G planes cols G columns G<br>Pool (dense grid) (smooth grid) (wavefct grid)<br>1 6 339 18817 6 339 18817 92 2668<br>2 6 339 18817 6 339 18817 92 2668<br>3 6 339 18817 6 339 18817 92 2668<br>4 6 338 18812 6 338 18812 92 2668<br>
5 6 338 18812 6 338 18812 93 2669<br>6 6 338 18812 6 338 18812 93 2669<br>7 6 338 18812 6 338 18812 93 2669<br>8 6 338 18812 6 338 18812 93 2669<br>9 6 338 18812 6 338 18812 93 2669<br>10 6 338 18812 6 338 18812 93 2669<br>
11 5 339 18817 5 339 18817 93 2669<br>12 5 339 18815 5 339 18815 93 2669<br>13 5 339 18815 5 339 18815 93 2669<br>14 5 339 18815 5 339 18815 92 2666<br>15 5 339 18815 5 339 18815 92 2666<br>16 5 339 18815 5 339 18815 92 2666<br>
tot 90 5417 301027 90 5417 301027 1481 42691<br>...<br>G cutoff = 1729.8995 ( 301027 G-vectors) FFT grid: ( 90, 90, 90)<br><br>Largest allocated arrays est. size (Mb) dimensions<br>Kohn-Sham Wavefunctions 4.63 Mb ( 2373, 128)<br>
NL pseudopotentials 9.12 Mb ( 2373, 252)<br>Each V/rho on FFT grid 1.48 Mb ( 48600, 2)<br>Each G-vector array 0.14 Mb ( 18817)<br>G-vector shells 0.01 Mb ( 1370)<br>Largest temporary arrays est. size (Mb) dimensions<br>Auxiliary wavefunctions 18.54 Mb ( 2373, 512)<br>
Each subspace H/S matrix 4.00 Mb ( 512, 512)<br>Each <psi_i|beta_j> matrix 0.49 Mb ( 252, 128)<br>Arrays for rho mixing 5.93 Mb ( 48600, 8)<br><br>Initial potential from superposition of free atoms<br>16 total processes killed (some possibly by mpirun during cleanup)<br>