On Fri, 24 Oct 2008, Carl R LeBlond wrote: hello carl,
CL> Hello - I've seem to run into this cholesky error. I've attached my CL> input file below which when run on a Cray XT3 MPP gives the following CL> error (pasted below). I've tried a number of suggestions I found on there are a number of details missing to have a full assessment of the situation. how did you compile the code (compilers, libraries) and which version or Q-E? i remember that some versions of ACML have problems and also PGI is not exactly known as a reliable compiler... i presume you are running on the PSC machine, right? they have multiple compiler versions and ACLM versions selectable as "modules". their defaults are usually overly conservative. :-( CL> this forum, however none of them seem to work in my case. Even CL> switching from davidson to cg did not alleviate the problem. The CL> problem began to appear after I increased my k points from 4 4 1 1 1 1 CL> too 8 8 8 1 1 1. Any help would be greatly appreciated. Probably CL> something blatantly obvious to a seasoned veteran. i made a check on a different machine, and your input runs fine for me, so it is not an intrinsic problem of the input or the code, but more likely related to libraries and compiler. you may want to check out using disk_io='none' and wf_collect=.true. for optimal use of resources. also it looks as if you don't take advantage of the -npools flag. which should help speeding up things even more. the G-space parallelization is far less efficient than parallelizing over k-space (and you have a lot of k-points to parallelize over). cheers, axel. CL> thanks much! CL> Carl CL> CL> Carl LeBlond CL> Assistant Professor of Chemistry CL> Indiana University of PA CL> Indiana, PA 15705 CL> CL> CL> CL> ...... CL> G cutoff = 320.4806 ( 961271 G-vectors) FFT grid: CL> (108,108,240) CL> G cutoff = 160.2403 ( 339703 G-vectors) smooth grid: ( 72, CL> 72,180) CL> CL> Largest allocated arrays est. size (Mb) dimensions CL> Kohn-Sham Wavefunctions 7.00 Mb ( 5331, 86) CL> NL pseudopotentials 31.24 Mb ( 5331, 384) CL> Each V/rho on FFT grid 5.34 Mb ( 349920) CL> Each G-vector array 0.92 Mb ( 120159) CL> G-vector shells 0.88 Mb ( 115730) CL> Largest temporary arrays est. size (Mb) dimensions CL> Auxiliary wavefunctions 55.97 Mb ( 5331, 688) CL> Each subspace H/S matrix 7.22 Mb ( 688, 688) CL> Each <psi_i|beta_j> matrix 0.50 Mb ( 384, 86) CL> Arrays for rho mixing 42.71 Mb ( 349920, 8) CL> CL> Initial potential from superposition of free atoms CL> CL> starting charge 143.87569, renormalised to 144.00000 CL> CL> negative rho (up, down): 0.750E-04 0.000E+00 CL> Starting wfc are 192 atomic wfcs CL> CL> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% CL> from pzpotrf : error # 49 CL> problems computing cholesky decomposition CL> %%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%% CL> CL> stopping ... CL> CL> CL> CL> CL> CL> CL> #### CL> &CONTROL CL> title = 'CH4 + Al(111)' , CL> calculation = 'relax' , CL> pseudo_dir = '/usr/users/8/leblond/pseudo/', CL> outdir = '/scratcha/leblond/' , CL> prefix = 'Al111' , CL> disk_io = 'default' , CL> verbosity = 'default' , CL> tprnfor = .true. , CL> nstep = 200 , CL> / CL> &SYSTEM CL> ibrav = 0, CL> celldm(1) = 7.65339026, CL> nat = 48, CL> ntyp = 1, CL> ecutwfc = 27.0000000000 , CL> ecutrho = 216.0000000000 , CL> nosym = .true. , CL> occupations = 'smearing' , CL> degauss = 0.03 , CL> smearing = 'methfessel-paxton' , CL> / CL> &ELECTRONS CL> electron_maxstep = 100, CL> conv_thr = 1.0e-08 , CL> startingpot = 'atomic' , CL> startingwfc = 'atomic' , CL> mixing_mode = 'local-TF' , CL> mixing_beta = 0.300000000 , CL> mixing_ndim = 8, CL> diagonalization = 'david' , CL> diago_david_ndim = 8, CL> / CL> &IONS CL> ion_dynamics = 'damp' , CL> trust_radius_max = 0.8D0 , CL> trust_radius_ini = 0.5D0 , CL> / CL> CELL_PARAMETERS hexagonal CL> 2.8284271 0.0000000 0.0000000 CL> 1.4142136 2.4494897 0.0000000 CL> 2.8284271 0.8164966 -5.7735027 CL> ATOMIC_SPECIES CL> Al 26.9815 Al.pbe-rrkj.UPF CL> ATOMIC_POSITIONS angstrom CL> Al 0.000000000000000 0.000000000000000 CL> 0.000000000000000 0 0 0 CL> Al 2.863782463805517 0.000000000000000 CL> 0.000000000000000 0 0 0 CL> Al 5.727564927611034 0.000000000000000 CL> 0.000000000000000 0 0 0 CL> Al 8.591347391416551 0.000000000000000 CL> 0.000000000000000 0 0 0 CL> Al 1.431891231902759 2.480108364567967 CL> 0.000000000000000 0 0 0 CL> Al 4.295673695708276 2.480108364567967 CL> 0.000000000000000 0 0 0 CL> Al 7.159456159513793 2.480108364567967 CL> 0.000000000000000 0 0 0 CL> Al 10.023238623319310 2.480108364567967 CL> 0.000000000000000 0 0 0 CL> Al 2.863782463805517 4.960216729135935 CL> 0.000000000000000 0 0 0 CL> Al 5.727564927611034 4.960216729135935 CL> 0.000000000000000 0 0 0 CL> Al 8.591347391416551 4.960216729135935 CL> 0.000000000000000 0 0 0 CL> Al 11.455129855222069 4.960216729135935 CL> 0.000000000000000 0 0 0 CL> Al 4.295673695708276 7.440325093703902 CL> 0.000000000000000 0 0 0 CL> Al 7.159456159513793 7.440325093703902 CL> 0.000000000000000 0 0 0 CL> Al 10.023238623319310 7.440325093703902 CL> 0.000000000000000 0 0 0 CL> Al 12.887021087124827 7.440325093703902 CL> 0.000000000000000 0 0 0 CL> Al 1.431891231902759 0.826702788189322 CL> -2.338268590217984 0 0 0 CL> Al 4.295673695708276 0.826702788189322 CL> -2.338268590217984 0 0 0 CL> Al 7.159456159513793 0.826702788189322 CL> -2.338268590217984 0 0 0 CL> Al 10.023238623319310 0.826702788189322 CL> -2.338268590217984 0 0 0 CL> Al 2.863782463805517 3.306811152757290 CL> -2.338268590217984 0 0 0 CL> Al 5.727564927611034 3.306811152757290 CL> -2.338268590217984 0 0 0 CL> Al 8.591347391416551 3.306811152757290 CL> -2.338268590217984 0 0 0 CL> Al 11.455129855222069 3.306811152757290 CL> -2.338268590217984 0 0 0 CL> Al 4.295673695708276 5.786919517325257 CL> -2.338268590217984 0 0 0 CL> Al 7.159456159513793 5.786919517325257 CL> -2.338268590217984 0 0 0 CL> Al 10.023238623319310 5.786919517325257 CL> -2.338268590217984 0 0 0 CL> Al 12.887021087124827 5.786919517325257 CL> -2.338268590217984 0 0 0 CL> Al 5.727564927611034 8.267027881893224 CL> -2.338268590217984 0 0 0 CL> Al 8.591347391416551 8.267027881893224 CL> -2.338268590217984 0 0 0 CL> Al 11.455129855222069 8.267027881893224 CL> -2.338268590217984 0 0 0 CL> Al 14.318912319027586 8.267027881893224 CL> -2.338268590217984 0 0 0 CL> Al 2.863782463805517 1.653405576378645 CL> -4.676537180435969 1 1 1 CL> Al 5.727564927611034 1.653405576378645 CL> -4.676537180435969 1 1 1 CL> Al 8.591347391416551 1.653405576378645 CL> -4.676537180435969 1 1 1 CL> Al 11.455129855222069 1.653405576378645 CL> -4.676537180435969 1 1 1 CL> Al 4.295673695708276 4.133513940946612 CL> -4.676537180435969 1 1 1 CL> Al 7.159456159513793 4.133513940946612 CL> -4.676537180435969 1 1 1 CL> Al 10.023238623319310 4.133513940946612 CL> -4.676537180435969 1 1 1 CL> Al 12.887021087124827 4.133513940946612 CL> -4.676537180435969 1 1 1 CL> Al 5.727564927611034 6.613622305514579 CL> -4.676537180435969 1 1 1 CL> Al 8.591347391416551 6.613622305514579 CL> -4.676537180435969 1 1 1 CL> Al 11.455129855222069 6.613622305514579 CL> -4.676537180435969 1 1 1 CL> Al 14.318912319027586 6.613622305514579 CL> -4.676537180435969 1 1 1 CL> Al 7.159456159513793 9.093730670082547 CL> -4.676537180435969 1 1 1 CL> Al 10.023238623319310 9.093730670082547 CL> -4.676537180435969 1 1 1 CL> Al 12.887021087124827 9.093730670082547 CL> -4.676537180435969 1 1 1 CL> Al 15.750803550930344 9.093730670082547 CL> -4.676537180435969 1 1 1 CL> K_POINTS automatic CL> 8 8 8 1 1 1 CL> ###### CL> _______________________________________________ CL> Pw_forum mailing list CL> Pw_forum at pwscf.org CL> http://www.democritos.it/mailman/listinfo/pw_forum CL> -- ======================================================================= Axel Kohlmeyer akohlmey at cmm.chem.upenn.edu http://www.cmm.upenn.edu Center for Molecular Modeling -- University of Pennsylvania Department of Chemistry, 231 S.34th Street, Philadelphia, PA 19104-6323 tel: 1-215-898-1582, fax: 1-215-573-6233, office-tel: 1-215-898-5425 ======================================================================= If you make something idiot-proof, the universe creates a better idiot.
