docs: add documentation for installation requirements of DPA-2

deepmd/pt/model/descriptor/repformers.py

@@ -60,7 +60,8 @@ def border_op(
         argument8,
     ) -> torch.Tensor:
         raise NotImplementedError(
-            "border_op is not available since customized PyTorch OP library is not built when freezing the model."
+            "border_op is not available since customized PyTorch OP library is not built when freezing the model. "
+            "See documentation for DPA-2 for details."
         )
 
     # Note: this hack cannot actually save a model that can be runned using LAMMPS.

doc/model/dpa2.md

@@ -8,6 +8,16 @@ The DPA-2 model implementation. See https://arxiv.org/abs/2312.15492 for more de
 
 Training example: `examples/water/dpa2/input_torch_medium.json`, see [README](../../examples/water/dpa2/README.md) for inputs in different levels.
 
+## Requirements of installation {{ pytorch_icon }}
+
+If one wants to run the DPA-2 model on LAMMPS, the customized OP library for the Python interface must be installed when [freezing the model](../freeze/freeze.md).
+
+The customized OP library for the Python interface can be installed by setting environment variable {envvar}`DP_ENABLE_PYTORCH` to `1` during installation.
+
+If one runs LAMMPS with MPI, the customized OP library for the C++ interface should be compiled against the same MPI library as the runtime MPI.
+If one runs LAMMPS with MPI and CUDA devices, it is recommended to compile the customized OP library for the C++ interface with a [CUDA-Aware MPI](https://developer.nvidia.com/mpi-solutions-gpus) library and CUDA,
+otherwise the communication between GPU cards falls back to the slower CPU implementation.
+
 ## Data format
 
 DPA-2 supports both the [standard data format](../data/system.md) and the [mixed type data format](../data/system.md#mixed-type).