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

How can I extrapolate a nonlinear curve to D->Inf limit?
by adfws 27 Sep '18

27 Sep '18

Hi everyone, I'm reading DMRG tutorials in http://alps.comp-phys.org/mediawiki/index.php/ALPS_2_Tutorials:DMRG-01_DMRG. and http://alps.comp-phys.org/mediawiki/index.php/ALPS_2_Tutorials:DMRG-02_Gaps It suggested me to extrapolate energies and energy gaps to D(maxstates) -> infinity. The question is about nonlinear extrapolation. In the case spin 1/2 gap versus 1/L, there is a linear relation and I can easily polyfit it linearly and get the gap in L->Inf limit. But in the case of quantities such as gaps versus D(maxstates), there is no simple linear relation. For example: Each time I enlarge D by 50, energy gaps change by the following values: [ -2.49851713e-08, -1.45270462e-11, -2.84217094e-14, -5.50670620e-14] which slows down changing by smaller and smaller values. I have done some exponential curve fittings in matlab but it become confusing when you have many models to fit it. How can I extrapolate such a nonlinear curve to D->Inf limit? Is there any standard way to do this? LiuBiao ll070616(a)126.com

2 1

Why spin 1 chain is added two spin 1/2 sites at its both ends?
by adfws 20 Sep '18

20 Sep '18

Hi everyone, I'm reading DMRG tutorials in http://alps.comp-phys.org/mediawiki/index.php/ALPS_2_Tutorials:DMRG-01_DMRG. When treating spin 1 Heisenberg chain with open boundary conditions, 2 spin 1/2 sites were added to both ends of the spin 1 chain. Why? I have completely no idea about this special treatment. Can anyone give further explanations? LiuBiao ll070616(a)126.com

2 1

Staggered susceptibility in dirloop_sse of a spin chain
by Stanislaw Galeski 07 Sep '18

07 Sep '18

Dear All, We are trying to reproduce results from this paper using the dirloop_sse code. https://journals.aps.org/prl/abstract/10.1103/PhysRevLett.89.047202 Typically we setup the simulation such that it calculates the spin-spin correlations and then calculate the staggered susceptibility from that. The results for uniform susceptibility obtained that way seem to reproduce well the known results, however it is not the case of the staggered susceptibility. In particular the staggered susceptibility seems to be the same for even and odd chain lengths and to scale in a bizarre way with length. This happens even for very short chains where we would expect the difference between odd/even to be significant. Is this a known problem with a simple explanation? Regards, Stanislaw and Kirill

2 4

The specific heat per site different for spinmc and qwl method!!!!!
by S Baidya 31 Aug '18

31 Aug '18

Dear ALPS users, I calculated specific heat capacity for square lattice with similar parameters, for spinmc (classical monte carlo) and qwl (qmc) for a square lattice. For spinmc, #prepare the input parameters parms = [] for l in [40]: for t in [5.0,4.5,4.0,3.5,3.0,2.9,2.8,2.7]: parms.append( { 'LATTICE' : "square lattice", 'T' : t, 'S' : 0.5, 'J' : 1 , 'THERMALIZATION' : 1000, 'SWEEPS' : 400000, 'UPDATE' : "cluster", 'MODEL' : "Ising", 'L' : l } ) for t in [2.6, 2.5, 2.4, 2.3, 2.2, 2.1, 2.0, 1.9, 1.8, 1.7, 1.6, 1.5, 1.2]: parms.append( { 'LATTICE' : "square lattice", 'T' : t, 'S' : 0.5, 'J' : 1 , 'THERMALIZATION' : 1000, 'SWEEPS' : 40000, 'UPDATE' : "cluster", 'MODEL' : "Ising", 'L' : l } ) #write the input file and run the simulation input_file = pyalps.writeInputFiles('parm7a',parms) pyalps.runApplication('spinmc',input_file,Tmin=5) For qwl, #prepare the input parameters parms = [{ 'LATTICE' : "square lattice", 'MODEL' : "spin", 'local_S' : 0.5, 'L' : 40, 'J' : -1 , 'Jxy' : 0, 'CUTOFF' : 1000 }] #write the input file and run the simulation input_file = pyalps.writeInputFiles('parm6a',parms) res = pyalps.runApplication('qwl',input_file) #run the evaluation and load all the plots data = pyalps.evaluateQWL(pyalps.getResultFiles(prefix='parm6a'),DELTA_T=0.1, T_MIN=0.1, T_MAX=10.0) I hoped that the two results should not show completely different specific heat per site curve with temperature. But, they are totally different. Can anyone please tell me if it is expected that classical and quantum result are completely different. Thanking you, Santu Baidya

3 4

Interaction involving three lattice sites
by Julien DESPRES 30 Aug '18

30 Aug '18

Dear ALPS community, I was wondering if it is possible to define an interaction involving three different lattice sites (for a 1D lattice model to perform dmrg calculations). To clarify, an example would be to construct the term \sum_i n_i n_{i+1} n_{i-1}. I checked in the doc if anyone had already the same kind of problem but it seems not. I guess that the solution is to build a bond term in the definition of the Hamiltonian by specifying the source and targets and to modify accordingly the lattice. Or is there a trick to specify one source (i) and two targets (i+1, i-1) ? Thanks in advance, Julien Ps : Just realized that I forgot to modify the subject (short vacations ...) -- __________________________________ DESPRES Julien Ph.D. student Center for Theoretical Physics Quantum Matter Theory team Ecole Polytechnique 91128 Palaiseau Cedex France, EU Office : 00.10.06 (ground floor) Phone : +33 1 69 33 42 81 Laboratoire Charles Fabry Atom Optics group Institut d'Optique Graduate School 2 Avenue Augustin Fresnel 91127 Palaiseau Cedex France, EU Office : R2.29 (1st floor) Phone: +33 1 64 53 33 49 __________________________________

2 1

What is the factor in Temperature unit when doing thermalization
by S Baidya 14 Aug '18

14 Aug '18

Dear Alps users, I tried to look into the examples and files in the alps directory to find the actual factor of temperature variable (T). In magnetization vs temperature plot in the example number mc-07 the plot shows temperature range of 0 to 5 in Kelvin. I need to know if the T is exactly T or with same factor. Where can I find the program to specify this details. Sometime T should be around 300-4000 K kelvin range but it still shows temperature around 5 - 8 Kelvin . Thank you, Santu Baidya

2 6

ALPS code error with Kagome spin lattice to calculate susceptibility
by S Baidya 08 Aug '18

08 Aug '18

Dear Alps users, I am totally new to ALPS code using quantum monte carlo. I have started to use this code for few small purpose like magnetic suceptibility vs temperature for various spin lattice models. Currently I have used ALPS code for calculating susceptibility for ising type S=3/2 Kagome lattice. I found in the lattice library that Kagome lattice is defined. So in the tutorial mc-02-susceptibilities/ I changed tutorial2d.py by editing #prepare the input parameters parms = [] for t in [0.05, 0.1, 0.2, 0.3, 0.4, 0.5, 0.6, 0.7, 0.8, 0.9, 1.0, 1.25, 1.5, 1.75, 2.0]: parms.append( { 'LATTICE' : "Kagome lattice", 'MODEL' : "spin", 'local_S' : 1.5, 'T' : t, 'J' : 4.2 , 'THERMALIZATION' : 5000, 'SWEEPS' : 50000, 'L' : 60, 'ALGORITHM' : "loop" } ) Then I changed the same way param2d file and ran the alpspython tutorial2d.py and I got strange error /usr/local/lib/python2.7/dist-packages/matplotlib/style/core.py:203: UserWarning: In /usr/share/matplotlib/stylelib/classic.mplstyle: text.dvipnghack is obsolete. Please remove it from your matplotlibrc and/or style files. warnings.warn(message) loop parm2d.in.xml ALPS/looper version 3.2b12-20100128 (2010/01/28) multi-cluster quantum Monte Carlo algorithms for spin systems available from http://wistaria.comp-phys.org/alps-looper/ copyright (c) 1997-2010 by Synge Todo <wistaria(a)comp-phys.org> using ALPS/parapack scheduler a Monte Carlo scheduler for multiple-level parallelization copyright (c) 1997-2016 by Synge Todo <wistaria(a)comp-phys.org> based on the ALPS libraries version 2.3.0 available from http://alps.comp-phys.org/ copyright (c) 1994-2016 by the ALPS collaboration. Consult the web page for license details. For details see the publication: B. Bauer et al., J. Stat. Mech. (2011) P05001. [2018-Aug-03 14:02:51]: starting scheduler on cces-System-Product-Name master input file = /home/cces/kagome-alps/mc-02/parm2d.in.xml master output file = /home/cces/kagome-alps/mc-02/parm2d.out.xml termination file = [disabled] total number of thread(s) = 1 thread(s) per clone = 1 number of thread group(s) = 1 auto evaluation = yes time limit = unlimited interval between checkpointing = 3600 seconds interval between progress report = 600 seconds interval between vmusage report = infinity task range = all worker dump format = hdf5 worker dump policy = running workers only [2018-Aug-03 14:02:51]: task status: total number of tasks = 15 new = 15, running = 0, continuing = 0, suspended = 0, finished = 0, completed = 0, skipped = 0 [2018-Aug-03 14:02:51]: starting 1 threadgroup(s) [2018-Aug-03 14:02:51]: dispatching a new clone[1,1] on threadgroup[1] WARNING: model is classically frustrated WARNING: model has negative signs I donot understand the actual reason and how to solve this situation. I used this tutorial before and it worked for ladder spin lattice. Can anyone please tell me what should I do to solve this situation. Thanking you, Santu Baidya

3 12

Problem in installing ALPS code in Ubuntu
by S Baidya 21 Jul '18

21 Jul '18

Dear ALPS users, I have downloaded alps-2.2.b4-src-with-boost.tar.gz in my linux pc with Ubuntu. I installed required python packages python2.7. I then untar by tar -xvzf alps-2.2.b4-src-with-boost.tar.gz Then mkdir build Then cd build Then cmake -DCMAKE_INSTALL_PREFIX=/home/cces/Downloads/ALPS/alps-2.2.b4-src-with-boost/alps/ .. I got an error. I tried different ways too using conda install but it did not work. So, I thought to build from source in linux as suggested in some websites. I have attached the Log file of the error. Can anyone please help me to install the alps in linux pc. I sent this mail before but I found there was an issue with maximum size of mail. So I have attached only one error file. Thanking you, Santu

4 20

vispython not found in recently installed alps-2.3 in Ubuntu !
by S Baidya 20 Jul '18

20 Jul '18

Dear alps users, After struggling I just installed alps-2.3 in Ubuntu . I build the alps-2.3 successfully. Now i was trying to do some tutorials in the directory ./alps-2.3.0-src-with-boost/alps/tutorials by runing the executable ./runtutorial.sh which fails saying /runtutorial.sh: 10: ./runtutorial.sh: vispython: not found ./runtutorial.sh: 10: ./runtutorial.sh: vispython: not found What should I do. Is there anything else needed to install to install alps successfully and use it for application. Please tell me to solution to use alps without error. Thanks, Santu

1 0

Reducing the number of correlation calculations in a run?
by PRADEEP 02 Jul '18

02 Jul '18

Dear ALPS community, I am using the DMRG programs (with OBC) to calculate the spin-spin correlations of a spin chain. Suppose N is the number of sites in the chain. My question is - while using the DMRG programs, is it possible to restrict calculations of correlations to a distance "d" between two sites where d<<N? Eg. instead of calculating all the N**2 correlations between all the spins, calculate only nearest neighbours or N-1 correlations? I want to obtain correlations for N>128 sites. ~60GB of RAM on a workstation is proving insufficient for this. I have been able to obtain local spin components <S_i^a> (a=x,y,z) for N=512 too. But, for the correlation calculations, N**2 correlations require more than 4 times the available RAM even for N=128. So I have to use smaller N, which leads to pronounced effects due to finite size and boundaries. Hence, the question. Thanks very much. with regards, Pradeep Thakur.

1 0