ambit
ambit copied to clipboard
jturney
C++ library for the implementation of tensor product calculations through a clean, concise user interface.
It is installed into ```lib/ambit/__init__.py``` when it should be installed into a sitelibdir. For example, ```/usr/local/lib/python3.8/site-packages```
This PR will add the ability to perform contractions using the TBLIS library. TBLIS is enabled at compilation by passing the arguments: ``` -DENABLE_TBLIS=ON -DTBLIS= ``` - [x] Add TBLIS...
This PR introduces a pool of temporary tensors to avoid repeated allocation/deallocation of intermediate tensors. Work in progress... - [ ] ready to go
This PR optimizes the algorithm to choose the best contraction permutation. # Summary ## - [x] Corrected error in computation of `cpu_cost`. - [ ] Modify the algorithm to consider...
Revision 9c7049a has these warnings: ``` /usr/ports/math/ambit/work/ambit-0.2-37-g9c7049a/src/tensor/print.cc:63:24: warning: passing an object of reference type to 'va_start' has undefined behavior [-Wvarargs] va_start(args, format); ^ /usr/ports/math/ambit/work/ambit-0.2-37-g9c7049a/src/tensor/print.cc:58:26: note: parameter of type 'const string...
It seems that ENABLE_TESTS should be a condition for this section? ``` #if (NOT ENABLE_PSI4) # test suite enable_testing() include(CTest) add_subdirectory(test) # sample suite add_subdirectory(samples) #endif() ```
Version 0.2 fails this way: ``` -- checking Fortran linkage: DGEMM_ -- checking Fortran linkage: dgemm_ -- checking Fortran linkage: DGEMM -- checking Fortran linkage: dgemm CMake Error at cmake/GenericMacros.cmake:169...
Our Python abilities are realized in two parts: Boost Python wrapping of C++ code and Python source code for operator overloading. The Boost Python module is compiled in the objdir...
## Improvements The following are improvements to Ambit that were brought up during the Tensor discussion at Emory with Psi4 developers: - [ ] Batching blocks/slices to/from disk - [...
The DF-MP2 example as per our discussion. ``` python # Initialize base objects iaQ = DiskTensor("iaQ DF Tensor", oshape, vshape, Qshape) eps_o = CoreTensor("Occupied eigenvalues", oshape) eps_v = CoreTensor("Virtual eigenvalues",...
