2017-07-05

• 1.4 alpha
• Stable release 1.3.3
• Changelog
• Documentation (PDF)
• Installation
• Features

• Prerequisites

  • Cmake 2.8 or higher
  • Python 2.6, 2.7, 3.2, 3.3, 3.4
  • Numpy 1.8.0 or higher
  • Scipy 0.10 or higher (0.12.0 or higher for python 3.3, 3.4)
  • h5py 2.3.0 or higher (requires HDF5 1.8.4 or higher)

• Compile core module

    cd pyscf/lib
    mkdir build; cd build
    cmake ..
    make
  

• To make python find pyscf, edit environment variable PYTHONPATH, e.g. if pyscf is installed in /opt, your PYTHONPATH should be

    export PYTHONPATH=/opt/pyscf:$PYTHONPATH
  

• Use Intel MKL as BLAS library. cmake with options -DBLA_VENDOR=Intel10_64lp_seq

    cmake -DBLA_VENDOR=Intel10_64lp_seq ..
  

  If cmake does not find MKL, you can define BLAS_LIBRARIES in CMakeLists.txt

    set(BLAS_LIBRARIES "${BLAS_LIBRARIES};/path/to/mkl/lib/intel64/libmkl_intel_lp64.so")
    set(BLAS_LIBRARIES "${BLAS_LIBRARIES};/path/to/mkl/lib/intel64/libmkl_sequential.so")
    set(BLAS_LIBRARIES "${BLAS_LIBRARIES};/path/to/mkl/lib/intel64/libmkl_core.so")
    set(BLAS_LIBRARIES "${BLAS_LIBRARIES};/path/to/mkl/lib/intel64/libmkl_avx.so")
  

• Using DMRG as the FCI solver for CASSCF. There are two DMRG solver interfaces avaialbe in pyscf. Block (http://chemists.princeton.edu/chan/software/block-code-for-dmrg) CheMPS2 (https://github.com/SebWouters/CheMPS2) After installing the DMRG solver, create a file future/dmrgscf/settings.py to store the path where the DMRG solver was installed.

• Using FCIQMC as the FCI solver for CASSCF. NECI (https://github.com/ghb24/NECI_STABLE) After installing the NECI, create a file future/fciqmc/settings.py to store the path where the NECI was installed.

• Using optimized integral library on X86 platform. Qcint (https://github.com/sunqm/qcint.git) is a branch of libcint library. It is heavily optimized against X86_64 platforms. To replace the default libcint library with qcint library, edit the URL of the integral library in lib/CMakeLists.txt file

    ExternalProject_Add(libcint
      GIT_REPOSITORY https://github.com/sunqm/qcint.git
      ...
  

• Error message "Library not loaded: libcint.2.9.dylib" On OS X. libcint.dylib is installed in pyscf/lib/deps/lib by default. Add "/path/to/pyscf/lib/deps/lib" to DYLD_LIBRARY_PATH

• On Mac OSX, error message of "import pyscf"

  OSError: dlopen(xxx/pyscf/lib/libcgto.dylib, 6): Library not loaded: libcint.2.8.dylib
  Referenced from: xxx/pyscf/lib/libcgto.dylib
  Reason: unsafe use of relative rpath libcint.2.8.dylib in xxx/pyscf/lib/libao2mo.dylib with restricted binary

It is only observed on OSX 10.11. One solution is to manually modify the relative path to absolute path

    $ install_name_tool -change libcint.2.8.dylib xxx/pyscf/lib/deps/lib/libcint.2.8.dylib xxx/pyscf/lib/libcgto.dylib
    $ install_name_tool -change libcint.2.8.dylib xxx/pyscf/lib/deps/lib/libcint.2.8.dylib xxx/pyscf/lib/libcvhf.dylib
    $ install_name_tool -change libcint.2.8.dylib xxx/pyscf/lib/deps/lib/libcint.2.8.dylib xxx/pyscf/lib/libao2mo.dylib
    ...

Running script pyscf/lib/_runme_to_fix_dylib_osx10.11.sh can patch all required dylib files.

• Fails at runtime with error message

  OSError: ... mkl/lib/intel64/libmkl_avx.so: undefined symbol: ownLastTriangle_64fc

This problem relates to MKL v11.1 on intel64 architecture. Currently, there is no solution for the combination of Python + MKL 11.1 + AVX. You need either change to other MKL version (10.*, 11.0, 11.2) or disable mkl_avx:

    cmake -DBLA_VENDOR=Intel10_64lp_seq .. -DDISABLE_AVX=1

• Runtime error message

  MKL FATAL ERROR: Cannot load libmkl_avx.so or libmkl_def.so.

This is MKL 11.* bug for "dlopen" function. Preloading the two libraries works fine with most system:

  export LD_PRELOAD=$MKLROOT/lib/intel64/libmkl_def.so:$MKLROOT/lib/intel64/libmkl_core.so

or

  export LD_PRELOAD=$MKLROOT/lib/intel64/libmkl_avx.so:$MKLROOT/lib/intel64/libmkl_core.so:$MKLROOT/lib/intel64/libmkl_sequential.so

• h5py installation. If you got problems to install the latest h5py package, you can try the old releases: https://www.hdfgroup.org/ftp/HDF5/releases/hdf5-1.8.12/ https://github.com/h5py/h5py/archive/2.3.1.tar.gz

The following paper should be cited in publications utilizing the PySCF program package:

• The Python-based Simulations of Chemistry Framework (PySCF), Q. Sun, T. C. Berkelbach, N. S. Blunt, G. H. Booth, S. Guo, Z. Li, J. Liu, J. McClain, E. R. Sayfutyarova, S. Sharma, S. Wouters, G. K.-L. Chan, arXiv:1701.08223v2 [physics.chem-ph]

Qiming Sun osirpt.sun@gmail.com