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

Problems about running python
by 赵炜 10 Aug '11

10 Aug '11

Dear Matthias or all users community, As is said in the tile, I have a problem about running python. It is that when I type "impot pyalps" and "import pyalps.plot", it appears on the screen that "ImportError: No module named scipy ". And I have no idea about that. Does anyone help,please? Thanks. zhaowei(赵炜）

2 1

ANDES/ALPS School on Numerical Methods for Many-Body Theories
by Matthias Troyer 04 Aug '11

04 Aug '11

Begin forwarded message: > From: "Adrian E. Feiguin" <afeiguin(a)uwyo.edu> > Date: 3 August 2011 17:30:28 MDT > > Dear colleague, > > We would like you to encourage students and young researchers to > participate in the following school on numerical methods to be held in > Bariloche, Argentina, Dec. 2-5 2011 and related conferences: > > ANDES/ALPS School on Numerical Methods for Many-Body Theories Dec. 2-5, 2011 > Bariloche, Argentina > > http://fisica.cab.cnea.gov.ar/mbt16/index.php/ANDES/ALPS_School_on_Numerica… > andhttp://fisica.cab.cnea.gov.ar/mbt16/ > > The deadline is pretty soon: August 5! > > We are attaching the flyer of both events and kindly ask you to please > print it out and hang it up in your group. > > We thank you for your collaboration, > > With best regards, > > > The ANDES/ALPS School Organizing Committee: > Pablo Cornaglia > Adrian Feiguin > Daniel Garcia > Karen Hallberg > Gonzalo Usaj > Web page:http://fisica.cab.cnea.gov.ar/mbt16/ > http://fisica.cab.cnea.gov.ar/mbt16/index.php/ANDES/ALPS_School_on_Numerica… > > Contact:mbt16@cab.cnea.gov.ar > > > > ------------------------------------------------------------------------------------------------------ > ANDES/ALPS School on Numerical Methods for Many-Body Theories > ------------------------------------------------------------------------------------------------------ > Bariloche, Argentina, Dec. 2 to 5, 2011 > > This school is aimed at advanced students in Physics and Chemistry. It > will cover several of the most important numerical methods used in > Many-Body Theories, taught by world leaders in the field. Some partial > support is available for a reduced number of students. There will be > hands-on activities for a practical learning of the computational > tools. > There is no registration fee for this activity. The activities will > take place at the campus of the Centro Atómico Bariloche (located > 9.5km West of the city of Bariloche) and students will be lodged on > campus and at hotels in Bariloche. > > LECTURERS: > Uli Schollwoeck: The Density Matrix Renormalization Group (DMRG) and > Matrix Product States > Adrián Feiguin: Time-dependent Methods > Simon Trebst: Series Expansions > Matthias Troyer: Quantum Monte Carlo > Philip Werner: Impurity Solvers for DMFT and Diagrammatic Monte Carlo Techniques > Antoine Georges, Michel Ferrero and Markus Aichhorn: Realistic > Simulations for Strongly Correlated Materials (DMFT+DFT) > Thomas Pruschke: The Numerical Renormalization Group (NRG) > > Deadlines: > Registration and Reception of Abstracts August 5, 2011 > Acceptance of Contributions August 26, 2011 > Registration fee (early payment) October 1, 2011 > (full support is available for a limited number of students upon request) > > Related conferences: > 1) 16th International Conference on Recent Progress in Many-Body > Theories - RPMBT16 > Bariloche, Argentina, Nov. 28 to Dec. 2 2011 > (http://fisica.cab.cnea.gov.ar/mbt16/) > > 2) New Frontiers in the Physics of Two Dimensional Electron Systems, > Buenos Aires, Nov. 23-25 2011 > (http://wwwtest.magnet.fsu.edu/mediacenter/seminars/icam/) > > -- > Adrian E. Feiguin > Assistant Professor > Dept. of Physics and Astronomy > University of Wyoming > 307-766-6534 > > >

1 0

Re: [ALPS-users] QMC simulation of time-dependent hamiltonian
by Matthias Troyer 03 Aug '11

03 Aug '11

Not for QMC and long times, since the real time evolution gives complex phases in the weights. Matthias On 3 Aug 2011, at 11:16, Dan Thompson wrote: > I think I understand what you mean now. It isn't possible to simulate a dynamic hamiltonian with alps. Ie., a MC time-dependent hamiltonian. > -Dan > > On Wed, Aug 3, 2011 at 1:39 PM, Matthias Troyer <troyer(a)phys.ethz.ch> wrote: > Hi Dan, > > That only makes sense if you are interested in the efficiency of our QMC algorithms. But the annealing time T in this scheme has nothing do to with real time. > > Matthias > > On 3 Aug 2011, at 10:37, Dan Thompson wrote: > >> Hey Matthias, >> >> Thanks for the quick reply. Here's what I'm trying to do with the alps qmc code: >> - Initially, I want to start with a quantum spin 1/2 system in a strong, transverse magnetic field. It should be in the ground state of this particular hamiltonian. >> - Then, the idea is to simulataneously: >> a) turn off the transverse field >> b) turn on a field along the spin axis and >> c) turn on couplings between the various spins. >> This would be the quantum annealing part and should take T MC time. >> - After the annealing, I would measure the new state of the spin system. Namely, I would look at the fidelity of this state and the ground state of the hamiltonian with the couplings and field along the spin axis. >> >> Does that make more sense? The idea would be to measure how the fidelity changes as the annealing time T changes. >> >> Cheers, >> -Dan >> >> >> >> On Wed, Aug 3, 2011 at 12:50 PM, Matthias Troyer <troyer(a)phys.ethz.ch> wrote: >> >> On 3 Aug 2011, at 09:40, Dan Thompson wrote: >> >> > Hello, >> > >> > I was just wondering if there was a way to simulate the quantum annealing of a spin chain/lattice. Namely, adjust the parameters in a model file so that over the course of the quantum monte carlo simulation I could start with one hamiltonian (a simple heisenberg hamiltonian with nearest neighbour couplings and a transverse magnetic field) and finish with a different one (transverse field turned off). >> > >> > Cheers, >> > -Dan >> >> Hi Dan, >> >> What is the purpose of this? Do you want to use it to improve the simulation or to get some time-dependent measurements? The QMC codes only measure equilibrium properties at the given parameters and you can thus just do a series of simulations. Please remember that MC time and real time are not the same. >> >> Matthias >> >> > >

1 0

Re: [ALPS-users] QMC simulation of time-dependent hamiltonian
by Matthias Troyer 03 Aug '11

03 Aug '11

Hi Dan, That only makes sense if you are interested in the efficiency of our QMC algorithms. But the annealing time T in this scheme has nothing do to with real time. Matthias On 3 Aug 2011, at 10:37, Dan Thompson wrote: > Hey Matthias, > > Thanks for the quick reply. Here's what I'm trying to do with the alps qmc code: > - Initially, I want to start with a quantum spin 1/2 system in a strong, transverse magnetic field. It should be in the ground state of this particular hamiltonian. > - Then, the idea is to simulataneously: > a) turn off the transverse field > b) turn on a field along the spin axis and > c) turn on couplings between the various spins. > This would be the quantum annealing part and should take T MC time. > - After the annealing, I would measure the new state of the spin system. Namely, I would look at the fidelity of this state and the ground state of the hamiltonian with the couplings and field along the spin axis. > > Does that make more sense? The idea would be to measure how the fidelity changes as the annealing time T changes. > > Cheers, > -Dan > > > > On Wed, Aug 3, 2011 at 12:50 PM, Matthias Troyer <troyer(a)phys.ethz.ch> wrote: > > On 3 Aug 2011, at 09:40, Dan Thompson wrote: > > > Hello, > > > > I was just wondering if there was a way to simulate the quantum annealing of a spin chain/lattice. Namely, adjust the parameters in a model file so that over the course of the quantum monte carlo simulation I could start with one hamiltonian (a simple heisenberg hamiltonian with nearest neighbour couplings and a transverse magnetic field) and finish with a different one (transverse field turned off). > > > > Cheers, > > -Dan > > Hi Dan, > > What is the purpose of this? Do you want to use it to improve the simulation or to get some time-dependent measurements? The QMC codes only measure equilibrium properties at the given parameters and you can thus just do a series of simulations. Please remember that MC time and real time are not the same. > > Matthias > >

1 0

QMC simulation of time-dependent hamiltonian
by Dan Thompson 03 Aug '11

03 Aug '11

Hello, I was just wondering if there was a way to simulate the quantum annealing of a spin chain/lattice. Namely, adjust the parameters in a model file so that over the course of the quantum monte carlo simulation I could start with one hamiltonian (a simple heisenberg hamiltonian with nearest neighbour couplings and a transverse magnetic field) and finish with a different one (transverse field turned off). Cheers, -Dan

2 1

Where is the measurements defined in hybridization dmft?
by Brains Chen 29 Jul '11

29 Jul '11

Dear all, I want to modify the measurement in the hybridization expansion of dmft and I looked through the source code in alps/applications/dmft/qmc and its two subdirectories hybridization/ and interaction_expansion/. In the hybridization/impurity.h, a paragraph of codes looks // create measurement objects measurements << RealVectorObservable("n"); measurements << RealVectorObservable("order"); measurements << vec_obs_t("Greens"); measurements << vec_obs_t("Greens_4point"); measurements << vec_obs_t("Greens_4point_ud"); measurements << RealObservable("sign"); measurements << RealObservable("MatrixSize"); measurements << RealVectorObservable("overlap"); Where is the 'measurements' defined? -- Kuang-Shing Chen

5 7

Quantum Monte Carlo for Ising in transverse field
by zhian asadzadeh 29 Jul '11

29 Jul '11

Dear Synge, > > no I don't have magnetic field in z direction, but looper results always > gives a finite Magnetization in the z direction. > > also I noticed that if we define the model like this > H=-J \sum_{i,j}S^{x}_{i}S^{x}_{j}-h \sum_{i}S^{z}_{i} > the looper doesn't work, I don't know why? > > Thanks, > Zhian. > Message: 3 Date: Wed, 27 Jul 2011 10:04:54 +0900 From: Synge Todo <wistaria(a)comp-phys.org> Subject: Re: [ALPS-users] Quantum Monte Carlo for Ising in transverse field To: comp-phys-alps-users(a)lists.phys.ethz.ch Message-ID: <17A32C75-B3B0-41ED-9F58-0304E71B8A3F(a)comp-phys.org> Content-Type: text/plain; charset=us-ascii On 2011/07/27, at 1:42, zhian asadzadeh wrote: > @Synge > I took it from Magnetization density, I think also Magnetization in z direction is not zero, > ok, thanks I will try that. Strange. Do you have a finite magnetic field along the z-axis, too? Synge

2 1

ALPS binary installer and MacOS X Lion
by Matthias Troyer 27 Jul '11

27 Jul '11

Dear ALPS users, The ALPS binary installers currently do not work on MacOS X Lion since Apple no longer includes the MPI libraries with MacOS X. You need to build ALPS from source on MacOS X Lion. We hope to release ALPS 2.0.2 soon, which will include support for Lion. Matthias

1 0

Quantum Monte Carlo for Ising in transverse field
by zhian asadzadeh 27 Jul '11

27 Jul '11

Dear Matthias and Synge I understood why there is that discrepancy for small clusters, it is true that we can write the thermal energy as : E(T)=\sum_{k} E(k)(2*n(k,T)-1) but this expression is true for thermodynamic limit, because we know the model has an even and odd parity sectors which they must be diagonalized separately, for T=0 the ground state is in even sector, so the sum over K points are for anti periodic boundary conditions, but for finite temperature we cant arrive to that expression, unless with this approximation that H(odd)=H(even) which is true in the thermodynamic limit, because the dispersion will be continuous and there is no much difference between odd and even space K points. for small sizes I checked the looper and exact diagonalization are same but with large difference with exact, but for large lattices the looper converges to exact result which is thermodynamic result: for L=200, h=0.3 looper:: -0.9238482(6.21e-05) with SWEEPS=4000000 exact:: -0.92385063 @Synge I took it from Magnetization density, I think also Magnetization in z direction is not zero, ok, thanks I will try that. Thanks, Zhian On Mon, Jul 25, 2011 at 12:00 PM, < 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. Quantum Monte Carlo for Ising in transverse field > (zhian asadzadeh) > 2. Re: Quantum Monte Carlo for Ising in transverse field > (Matthias Troyer) > 3. Re: Quantum Monte Carlo for Ising in transverse field (Synge Todo) > 4. Custom measurements (Alexander Herzog) > > > ---------------------------------------------------------------------- > > Message: 1 > Date: Sun, 24 Jul 2011 12:48:03 +0200 > From: zhian asadzadeh <zhian.asadzadeh(a)gmail.com> > Subject: [ALPS-users] Quantum Monte Carlo for Ising in transverse > field > To: comp-phys-alps-users(a)lists.phys.ethz.ch > Message-ID: > <CAEw6G6sEs2aaE8oY5S+dFob35edARq12xKp9=4h4BUdrZZ1i7g(a)mail.gmail.com > > > Content-Type: text/plain; charset="iso-8859-1" > > Dear Matthias, > > by \Sigma_{i,j}, I mean \sum_{i,j}, > > for the exact result, we know the model is integrable, so by fermionization > and finally diagonalizing the Hamiltonian we will end up to an Hamiltonian > with sum over independent modes, > H=\sum_{k} E(k) (2 a_{k}^{dagger} a_{k}-1) > > for the T=0 ground state energy, I have compared with DMRG results which > are > exactly same, so I am sure up to this point the form of E(k) and my > summation over k space is correct. > > so finally for E(T) simply we have: > E(T)=\sum_{k} E(k)(2*n(k,T)-1) in which n(k,T)=1/(exp(E(k)/T)+1) ::with > Boltzmann constant=1 > > > Regards, > Zhian. > > > On Sun, Jul 24, 2011 at 12:00 PM, < > 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. Quantum Monte Carlo for Ising in transverse field > > (zhian asadzadeh) > > 2. Re: Quantum Monte Carlo for Ising in transverse field > > (Matthias Troyer) > > > > > > ---------------------------------------------------------------------- > > > > Message: 1 > > Date: Sat, 23 Jul 2011 16:29:39 +0200 > > From: zhian asadzadeh <zhian.asadzadeh(a)gmail.com> > > Subject: [ALPS-users] Quantum Monte Carlo for Ising in transverse > > field > > To: comp-phys-alps-users(a)lists.phys.ethz.ch > > Message-ID: > > <CAEw6G6sPavSO0RMreyy701xo+LN4M4LGrBzOLJkonEKs9S= > Wkw(a)mail.gmail.com > > > > > Content-Type: text/plain; charset="iso-8859-1" > > > > Dear Matthias, > > > > my intention is doing finite T QMC for 2D ising in transverse field > with > > the following Hamiltonian: > > H=-J \Sigma_{i,j} S_{i}^{z} S_{j}^{z}-h \Sigma_{i} S_{i}^{x} > > but for checking the accuracy of the numeric I did some checks for 1D > > case which we can have an exact results, > > probably I am expecting too much accuracy, for example: > > > > L=100, T=0.32859 > > Exact:: E(T)=-105.29789708270407, Mx(T)=0.27030367598168420 > > looper:: E(T)=-105.61795 (0.00554), M(T)=0.4093185 (0.000138) ( I > > think it is in z direction) > > with SWEEPS:6000000, THERMALIZATION=2000000 > > > > I have checked also for other sizes and very large SWEEPS. > > I don't expect to get exact result but I need much more accuracy in > the > > calculated energy. > > > > and also other question: that magnetization which looper calculates > is > > in the z direction? > > can I also measure M_{x} just with declaring it in parm file? > > > > Thanks a lot, > > Zhian. > > > > > > On Sat, Jul 23, 2011 at 12:00 PM, < > > 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. restarting Lanczos iteration (jessica.alfonsi(a)unipd.it) > > > 2. Quantum Monte Carlo for Ising in transvers field (zhian asadzadeh) > > > 3. Re: Quantum Monte Carlo for Ising in transvers field > > > (Matthias Troyer) > > > 4. Re: restarting Lanczos iteration (Matthias Troyer) > > > > > > > > > ---------------------------------------------------------------------- > > > > > > Message: 1 > > > Date: Fri, 22 Jul 2011 12:22:47 +0200 > > > From: jessica.alfonsi(a)unipd.it > > > Subject: [ALPS-users] restarting Lanczos iteration > > > To: comp-phys-alps-users(a)lists.phys.ethz.ch > > > Message-ID: > > > <5b0ae17b0813c63a97720e0fbc6cc1c6.squirrel(a)webmail.unipd.it> > > > Content-Type: text/plain;charset=utf-8 > > > > > > Hi all, > > > I'd like to modify the Lanczos *.cpp examples in IETL example folder to > > be > > > able to restart the iteration from any interruption point during the > > > Lanczos iteration. I suppose first one has to dump some information, > for > > > instance the cofficients of T-matrix and the starting vector, however > the > > > more difficult part would be the resuming of the iteration. Can > somebody > > > give me some hints on how accomplishing this task? Thanks in advance... > > > > > > Best regards, > > > > > > Jessica Alfonsi > > > > > > > > > > > > > > > > > > ------------------------------ > > > > > > Message: 2 > > > Date: Fri, 22 Jul 2011 13:03:23 +0200 > > > From: zhian asadzadeh <zhian.asadzadeh(a)gmail.com> > > > Subject: [ALPS-users] Quantum Monte Carlo for Ising in transvers field > > > To: comp-phys-alps-users(a)lists.phys.ethz.ch > > > Message-ID: > > > < > > CAEw6G6vLtP2G8kviZ97cxgZy9Dv1x_MqyUzFWs9ejnM8WvwnmA(a)mail.gmail.com > > > > > > > Content-Type: text/plain; charset="iso-8859-1" > > > > > > Dear All, > > > > > > I am trying to do QMC at finite temperature for calculating the > > > magnetization of *Ising model in transverse field*, but I realized just > > > looper code is working for this model, ( the others codes doesn't work > > with > > > off diagonal matrix elements) but the result are not so good. > > > I am interested to modify dirloop_sse for this model, is it possible? > > > I appreciate your help, > > > Thanks, > > > zhian. > > >

3 2

TEBD: kink state problem
by juan florez 26 Jul '11

26 Jul '11

Dear Matthias Troyer and ALPS developers, I've been testing the TEBD routines by using python programs as the ones present in the tutorials. I wanted to go further from what we can find in such tutorials because of my research interests and according to the possibilities of the TEBD routines of the ALPS 2.0 illustrated in the paper: The ALPS project release 2.0: open source software for strongly correlated systems. What I just found is that if I use a general spin value in the spin chains (TEBD-02) let say S, the initial "kink" state of the chain does not represent a general down,down,....,up,up.... states for the system but just a partial spin flit is performed in the initial aligned state. I check the numerical results at the very initial moment of the spin dynamics calculation and the "kink" state for a S system turns out to be S S S......, S-0.5 S-0.5 S-0.5 ....... which means a 0.5 less in the second part of the vector state instead of being S S S......, -S -S -S as it should be. I read the ALPS web page and paper and I couldn't see anything expressing that the TEBD routines are just working for spin 1/2 in the case of the kink state. If it is so, how can I modify such a situation in order to work with different spin values?. The ALPS paper says that the TEBD routines should work for the "spin" models in the ALPS and such models allow modifications of the spin value. Thanks for your answer. JM.

1 0