Unable to run the hybridization tutorial scripts; cthyb and boost.mpi not found
Hello, Recently I found that an implementation of the continuous-time QMC hybridization scheme (CT-HYB) has been included in the ALPS library and I'd like to learn how it works, so I installed the binary packages of alps 2.2.0b3, vistrails 2.1.2 and alps-vistrails-2.2.0b3 on Mac OS X 10.10.3. I've tested a few tutorials to make sure the installation was complete. However, tutorial1.py in the hybridization tutorial told me that it couldn't find the cthyb module, and tutorial2~4 told me it couldn't find both boost.mpi and cthyb. Is it because the v2.2.0b3 binary package does not have them compiled and I have to rebuild the library from the source code? (I don't have MPI or Open MPI installed on my Mac, by the way. Does it matter?) Moreover, I tried to follow the instruction in doc/hybdoc.pdf to run hybridization from scratch, but there's a file "delta.dat" containing hybridization functions which I don't know how to produce. Did I miss something in the tutorial? Thank you. ps. I'm aware that CT-HYB is implemented as an impurity solver for the DMFT applications. But my project does not need DMFT calculations and I intend to simply use hybridization as a stand-alone module. Sincerely, Leo (Fang Yao-Lung) Duke Physics
Hi Leo,
Moreover, I tried to follow the instruction in doc/hybdoc.pdf to run hybridization from scratch, but there's a file "delta.dat" containing hybridization functions which I don't know how to produce. Did I miss something in the tutorial?
I can help you with that. For your application, the python scripts may not be the best way to proceed.
ps. I'm aware that CT-HYB is implemented as an impurity solver for the DMFT applications. But my project does not need DMFT calculations and I intend to simply use hybridization as a stand-alone module.
In order to simulate an Anderson impurity model you will need to have a hybridization function, parametrized in imaginary time. We can set one up if you’re interested, but ideally it comes from your physics setup. Best regards, Emanuel
Hello Prof. Gull,
Moreover, I tried to follow the instruction in doc/hybdoc.pdf to run hybridization from scratch, but there's a file "delta.dat" containing hybridization functions which I don't know how to produce. Did I miss something in the tutorial?
I can help you with that. For your application, the python scripts may not be the best way to proceed.
My intention was simply using the python scripts to check if I could reproduce Figs 1 & 2 in Comp. Phy. Comm. 184, 1280 (2013) for the single-orbital Anderson model as the first step, so tutorial hybridization-02 is essential to me. But I agree python scripts may not be the best way for my project (see below).
ps. I'm aware that CT-HYB is implemented as an impurity solver for the DMFT applications. But my project does not need DMFT calculations and I intend to simply use hybridization as a stand-alone module.
In order to simulate an Anderson impurity model you will need to have a hybridization function, parametrized in imaginary time. We can set one up if you’re interested, but ideally it comes from your physics setup.
I was looking for the format of this dat file but it seems not mentioned in the documentation. However, I just noticed that a delta.dat file with constant hybridization function could be generated from the script of tutorial hybridization-01 (after skipping the cthyb and boost.mpi errors). I assume the first column is for the imaginary time index and the second and the third for spin up and down, respectively. Is it correct? By feeding this dat to the hybridization executable, I successfully finished a calculation with a bunch of dat file and a h5 file generated. My problem at hand involves one quantum dot coupled to two metal leads, so basically it's kind of like a single-orbital Anderson impurity model. Later I may need to increase the number of impurities and to include interactions with other physical degree of freedoms, which may complicate the update scheme. I guess eventually I may have to tweak the hybridization executable to fit my need. Thank you. Sincerely, Leo (Fang Yao-Lung) Duke Physics
I think the code will work perfectly for you. You should read the manual called ‘hybdoc.pdf’ which I will send you in a separate mail (you can also find it in qmc/hybridization/Documentation). Cheers, Emanuel
On May 7, 2015, at 3:44 PM, Leo Fang <leofang@phy.duke.edu> wrote:
Hello Prof. Gull,
Moreover, I tried to follow the instruction in doc/hybdoc.pdf to run hybridization from scratch, but there's a file "delta.dat" containing hybridization functions which I don't know how to produce. Did I miss something in the tutorial?
I can help you with that. For your application, the python scripts may not be the best way to proceed.
My intention was simply using the python scripts to check if I could reproduce Figs 1 & 2 in Comp. Phy. Comm. 184, 1280 (2013) for the single-orbital Anderson model as the first step, so tutorial hybridization-02 is essential to me. But I agree python scripts may not be the best way for my project (see below).
ps. I'm aware that CT-HYB is implemented as an impurity solver for the DMFT applications. But my project does not need DMFT calculations and I intend to simply use hybridization as a stand-alone module.
In order to simulate an Anderson impurity model you will need to have a hybridization function, parametrized in imaginary time. We can set one up if you’re interested, but ideally it comes from your physics setup.
I was looking for the format of this dat file but it seems not mentioned in the documentation. However, I just noticed that a delta.dat file with constant hybridization function could be generated from the script of tutorial hybridization-01 (after skipping the cthyb and boost.mpi errors). I assume the first column is for the imaginary time index and the second and the third for spin up and down, respectively. Is it correct? By feeding this dat to the hybridization executable, I successfully finished a calculation with a bunch of dat file and a h5 file generated.
My problem at hand involves one quantum dot coupled to two metal leads, so basically it's kind of like a single-orbital Anderson impurity model. Later I may need to increase the number of impurities and to include interactions with other physical degree of freedoms, which may complicate the update scheme. I guess eventually I may have to tweak the hybridization executable to fit my need.
Thank you.
Sincerely, Leo (Fang Yao-Lung) Duke Physics
participants (2)
-
Emanuel Gull
-
Leo Fang