NNPOps
NNPOps copied to clipboard
openmm
High-performance operations for neural network potentials
Compiling with CUDA 12 and a very recent pytorch version (such as v2.1.0 from the nightly) will make compilation fail because C++17 is required to compile pytorch: ```shell (test7) $...
It is required to have to build nnpops with version 11.* of cudatoolkit and a 10.3 gxx compile. These are going to be outdated. When I try to compile with...
The underlying CFConv implementation supports PBC, but CFConvNeighbors just hardcodes the box vectors as null. https://github.com/openmm/NNPOps/blob/d15cb9196e283b6b55f88a93d85232458f64fa18/src/pytorch/CFConvNeighbors.cpp#L74 @raimis I believe you wrote this, do you remember why it was decided not...
I'm trying to build NNPOps on a Mac. It builds correctly, but when I try to `import NNPOps` it fails with this error: ``` Traceback (most recent call last): File...
This PR is an effort to include the possibility of installing NNPops via pip. I mainly took the example here and adapted it: https://github.com/pybind/cmake_example/blob/master/setup.py Having pip as an option would...
Hi, when I simulate a 15 Angstrom waterbox with the `torchani` and `nnpops` implementation the `torchani` implementation is slightly faster. Is `nnpops` only outperforming `torchani` with small system size? I...
The index here overflows with 128K particles (a thread is launched per possible pair, N*(N-1)/2) https://github.com/openmm/NNPOps/blob/054d487d9fec8f98a111bc30ec6d3ec1ce423356/src/pytorch/neighbors/getNeighborPairsCUDA.cu#L43-L46 I know it sounds like an absurd regime for such a brute force approach,...
We could enforce some agreement on formatting throughout the code using an automatic tool, such as clang-format for C++ and [black](https://black.readthedocs.io/en/stable/) for Python. A .clang-format at the root of a...
Hi all, Thanks very much for making this development effort public and modular! Is there a listing somewhere here of the primitive operations that are provided by NNPOps (not the...
If I try to enable building CUDA, CMake fails with this error: ``` CMake Error at CMakeLists.txt:8 (enable_language): Language 'CUDA' is currently being enabled. Recursive call not allowed. ``` I...