- Comp-phys-alps-users - lists.phys.ethz.ch

problems regarding the MPI runs in the ALPS
by D´Souza, Sunil Wilfred 12 Dec '14

12 Dec '14

Dear Alps users, I have installed ALPS2.2 on my local machine containing 12 processing units. I have tried running a simple tutorial given in the alps examples files. when i am using the following command "pyalps.runApplication('spinmc',input_file,Tmin=5)" , everything works fine and the ALPS program runs (Serial mode) and gives the proper output as expected. On the other hand when i use the command "pyalps.runApplication('spinmc',input_file,Tmin=5,MPI=7)" , i am getting a lot of error messages related to segmentation fault. The error messages are as follows: mpirun -np 7 spinmc --mpi parm2a.in.xml --Tmin 5 [LIN21:04682] *** Process received signal *** [LIN21:04682] Signal: Segmentation fault (11) [LIN21:04682] Signal code: Address not mapped (1) [LIN21:04682] Failing at address: (nil) [LIN21:04682] [ 0] /lib64/libpthread.so.0(+0xf9f0) [0x7f356ffcf9f0] [LIN21:04682] [ 1] /opt/openmpi/1.4.5/gcc/lib64/openmpi/mca_pml_v.so(+0x1b62) [0x7f3569f92b62] [LIN21:04682] [ 2] /opt/openmpi/1.4.5/gcc/lib64/libopen-pal.so.0(mca_base_components_close+0x76) [0x7f356f2616c6] [LIN21:04682] [ 3] /opt/openmpi/1.4.5/gcc/lib64/libmpi.so.0(mca_pml_base_select+0x301) [0x7f356f765d11] [LIN21:04682] [ 4] /opt/openmpi/1.4.5/gcc/lib64/libmpi.so.0(+0x39acc) [0x7f356f725acc] [LIN21:04682] [ 5] /opt/openmpi/1.4.5/gcc/lib64/libmpi.so.0(MPI_Init+0x16b) [0x7f356f745a8b] [LIN21:04682] [ 6] /home/sunil/ALPS2.2/lib/libalps.so.2(_ZN4alps9comm_initERiRPPcb+0x58) [0x7f3570f60098] [LIN21:04682] [ 7] /home/sunil/ALPS2.2/lib/libalps.so.2(_ZN4alps9scheduler5startEiPPcRKNS0_7FactoryE+0x37) [0x7f35711a08e7] [LIN21:04682] [ 8] spinmc(main+0x15) [0x449a95] [LIN21:04682] [ 9] /lib64/libc.so.6(__libc_start_main+0xf5) [0x7f356d184be5] [LIN21:04682] [10] spinmc() [0x449de5] [LIN21:04682] *** End of error message *** -------------------------------------------------------------------------- mpirun noticed that process rank 1 with PID 4682 on node LIN21 exited on signal 11 (Segmentation fault). -------------------------------------------------------------------------- I am very kindly requesting the ALPs users to help me out in solving the above mentioned issue. I am looking forward for your suggestions to overcome this problem. with best regards Sunil

2 1

Exact Diagonalization Fermion Hubbard
by Sriluckshmy Viswanathan 12 Dec '14

12 Dec '14

Dear all, I am trying to solve the fermion hubbard model on a square lattice with a small cluster size of 2x2 using exact diagonalization. I want to look at the energy levels at half filling. I used to N_total parameter to fix the filling, but the energy levels results dont change if I change this parameter. Could you please help me with this problem? I have attached the param file that I used. I am using fulldiag of ALPS. Regards Sriluckshmy

3 5

Re: [ALPS-users] Comp-phys-alps-users Digest, Vol 105, Issue 1
by Mateusz Łącki 01 Dec '14

01 Dec '14

ok, to weź poproś - chciałem Ci skopiować oprócz skryptów wyniki minimalizacji energii dla tych danych dla których om chciał wyciągnąć „Sz“. Nie miałem póki co czasu tego zrobić (bo Kuba chciał inne rzeczy „na wczoraj“, a także tutaj mam sporo czasochłonnych obowiązków). > On 01 Dec 2014, at 12:00, comp-phys-alps-users-request(a)lists.phys.ethz.ch wrote: > > Send Comp-phys-alps-users mailing list submissions to > comp-phys-alps-users(a)lists.phys.ethz.ch > > To subscribe or unsubscribe via the World Wide Web, visit > https://lists.phys.ethz.ch/listinfo/comp-phys-alps-users > or, via email, send a message with subject or body 'help' to > comp-phys-alps-users-request(a)lists.phys.ethz.ch > > You can reach the person managing the list at > comp-phys-alps-users-owner(a)lists.phys.ethz.ch > > When replying, please edit your Subject line so it is more specific > than "Re: Contents of Comp-phys-alps-users digest..." > > > Today's Topics: > > 1. Exact Diagonalization Fermion Hubbard (Sriluckshmy Viswanathan) > 2. Re: Exact Diagonalization Fermion Hubbard (Michele Dolfi) > 3. Re: Exact Diagonalization Fermion Hubbard > (Sriluckshmy Viswanathan) > 4. Re: bug in mps_optim (Michele Dolfi) > 5. Re: Exact Diagonalization Fermion Hubbard (Matthias Troyer) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Sun, 30 Nov 2014 16:54:43 +0530 > From: Sriluckshmy Viswanathan <sriluckus(a)gmail.com> > To: comp-phys-alps-users <comp-phys-alps-users(a)lists.phys.ethz.ch> > Subject: [ALPS-users] Exact Diagonalization Fermion Hubbard > Message-ID: > <CAJTRBgcmbU1Fz=zr+fPxRm9id6JAD66RPBYgn_JEouVkbbjuXw(a)mail.gmail.com> > Content-Type: text/plain; charset="iso-8859-1" > > Dear all, > > I am trying to solve the fermion hubbard model on a square lattice with a > small cluster size of 2x2 using exact diagonalization. I want to look at > the energy levels at half filling. I used to N_total parameter to fix the > filling, but the energy levels results dont change if I change this > parameter. Could you please help me with this problem? I have attached the > param file that I used. I am using fulldiag of ALPS. > > Regards > Sriluckshmy >

1 0

bug in mps_optim
by Dominique Delande 30 Nov '14

30 Nov '14

Dear all, I think that there is a bug in how mps_optim changes the bond dimension during successive sweeps. In the alps-2.2.b3-r7462-src/alps/applications/dmrg/mps/mps_optim/run_sim.cpp file, line 64 reads: double step = (int(parms["MAXSTATES"]) - init_bond) / (parms["nsweeps"]-1); It is supposed to compute the increment of the bond size at each step (starting from init_bond at the first sweep, reaching MAXSTATES at the final sweep). How, the division if performed using integers before being written to the variable step as a double. For example, if (int(parms["MAXSTATES"]) - init_bond) < (parms["nsweeps"]-1), step will be 0.0, presumably not the intended value. A simple fix is to perform the division using double, something like: double step = ((double)(int(parms["MAXSTATES"]) - init_bond)) / (parms["nsweeps"]-1); Best regards Dominique Delande -- Dominique Delande (Dominique.Delande(a)lkb.upmc.fr) Laboratoire Kastler-Brossel - Case 74 - Universite P. et M. Curie 4, place Jussieu, F-75252 Paris Cedex 05, FRANCE Phone : +33 1 44 27 27 97 - Fax : +33 1 44 27 38 45 Acces : Tour 13, Couloir 12-13, 3eme etage - Bureau 316

2 1

Re: [ALPS-users] Results is different for worm(sse) and wolfram mathematica
by vitbka konef 25 Nov '14

25 Nov '14

My name is Konev Vitaliy. 1) Yes , it does. 2) In Mathematica I use the periodic boundary conditions. How realised the periodic boundary conditions in ALPS library? I was attached test of the result for periodic boundary condition.

2 3

Results is different for worm(sse) and wolfram mathematica
by vitbka konef 22 Nov '14

22 Nov '14

First , I was obtained density of magnetization and density of energy curves of isotropic Heisenberg model on square lattice (2x2) using worm and dirloop_sse. After I was obtained these curves with wolfram mathematica. Curves for magnetization density(density of energy) obtained with mathematica and worm(sse) is strongly differ one from another.They must match. I was attached the files of mathematica and Heis2x2.py.

2 1

total acceptance rate
by tesfaye mamuye 20 Nov '14

20 Nov '14

Dear ALPS developers and users I'm using CT-HYB impurity solver for single site DMFT problem. While i was simulating single site single-orbital dmft problem using CT-HYB the value of total acceptance rate negative. Is this negative value of total acceptance rate possibly accepted ?what is this negative value of total acceptance rate mean ? Tesfaye MamuyeAddis Ababa University, Ethiopia

1 0

Anisotropic XY model in transverse magnetic field
by Олеся Крупніцька 17 Nov '14

17 Nov '14

Dear ALPS team, I want to inform you that I have found my mistake in extracting results for magnetization of anisotropic XY model on frustrated diamond spin chain in transverse magnetic field. Best regards, Olesia Krupnitska

1 0

Anisotropic XY model in transverse magnetic field
by Олеся Крупніцька 13 Nov '14

13 Nov '14

Dear ALPS team, I am trying to obtain magnetization curve of anisotropic XY model on frustrated diamond spin chain in transverse magnetic field (m^x) using full diagonalization. For this purpose I have prepared such input file for delta=1.5; LATTICE="lattice_diamond" MODEL="spin" LATTICE_LIBRARY="lattice-diamond.xml" local_S=1/2 Jxy1 = 0.85 Jxy2 = 3.0 Jxy3 = 1.15 Jz1 = 1.275 Jz2 = 4.5 Jz3 = 1.725 L = 2 T = 0.001 {Gamma =4;} {Gamma =4.05;} {Gamma =4.1;} {Gamma =4.15;} {Gamma =4.2;} {Gamma =4.25;} {Gamma =4.3;} {Gamma =4.35;} {Gamma =4.4;} {Gamma =4.45;} {Gamma =4.5;} {Gamma =4.55;} {Gamma =4.6;} {Gamma =4.65;} {Gamma =4.7;} {Gamma =4.75;} {Gamma =4.8;} {Gamma =4.85;} {Gamma =4.9;} {Gamma =4.95;} {Gamma =5;} {Gamma =5.05;} {Gamma =5.1;} {Gamma =5.15;} {Gamma =5.2;} {Gamma =5.25;} {Gamma =5.3;} {Gamma =5.35;} {Gamma =5.4;} {Gamma =5.45;} {Gamma =5.5;} {Gamma =5.55;} {Gamma =5.6;} {Gamma =5.65;} {Gamma =5.7;} {Gamma =5.75;} {Gamma =5.8;} {Gamma =5.85;} {Gamma =5.9;} {Gamma =5.95;} {Gamma =6;} lattice-diamond.xml is: <LATTICES> <LATTICEGRAPH name="lattice_diamond"> <FINITELATTICE> <LATTICE dimension="1"> <BASIS> <VECTOR>1</VECTOR> </BASIS> </LATTICE> <PARAMETER name="L"/> <EXTENT size="L"/> <BOUNDARY type="periodic"/> </FINITELATTICE> <UNITCELL dimension="1" vertices="3"> <VERTEX id="1" type="0"> </VERTEX> <VERTEX id="2" type="0"> </VERTEX> <VERTEX id="3" type="0"> </VERTEX> <EDGE type="2"> <SOURCE vertex="1"/> <TARGET vertex="2"/> </EDGE> <EDGE type="3"> <SOURCE vertex="1"/> <TARGET vertex="3"/> </EDGE> <EDGE type="1"> <SOURCE vertex="2"/> <TARGET vertex="3"/> </EDGE> <EDGE type="3"> <SOURCE vertex="3"/> <TARGET vertex="2" offset="1"/> </EDGE> <EDGE type="1"> <SOURCE vertex="3"/> <TARGET vertex="1" offset="1"/> </EDGE> </UNITCELL> </LATTICEGRAPH> </LATTICES> It works perfect for L=2 (6 sites), but I cant obtain correct magnetization curve for L=3 (9 sites) or L= 4 (12 sites) etc. Would you be so kind to explain me why it works well for small chain and worse for slightly larger chain? Best regards, Olesia Krupnitska

1 0

correlation function and alternating chain
by Jalal Jahangiri 01 Nov '14

01 Nov '14

Hello Dear ALPs users I have 2 questions.I am novice in ALPS. 1- I want to calculate correlation function for coupled ladders. my input file is attached in the output file, there are 4 values of correlation function for nearest neighbors. this is part of output file *Spin Correlations*[( 0,0 ) -- ( 0,1 )]*Spin Correlations*[( 0,0 ) -- ( 0,1 )]*Spin Correlations*[( 0,0 ) -- ( 0,1 )]*Spin Correlations*[( 0,0 ) -- ( 0,1 )] -whereas i expect there is one of them or at least 3 of them (x-y-z), is there a bug in the input file or something else? 2- How do I define an alternating chain? (my mean is alternating coupling constants) Best Regards

1 0