2025-Feb09 main code

Updates:

1. MC moves refactored into move classes
2. small changes for gRASPA_pybind

Jan13-2025 Release

small bug fix (writing movie files)

2024-Dec12 main code

Dec. 2024 backup release

2024-Nov05 main code

some speedup, 10% - 20%
excess loading for GCMC when using Peng-Robinson EOS

2024-Sept30 main code

end of Sept. backup release
some warning fixes

2024-Sept12 main code

PR: heat of adsorption for single and multiple components

heat of adsorption for single and multiple components
with new XeKr mixture, Bae mix-ligand CO2/methane mixture, and NU2000 p-xylene tests

2024-Aug12 main code

main gRASPA code @ 081224

gRASPA-HTC: high-throughput-calculations

A standalone version of the gRASPA code that runs a GCMC simulation per CUDA block.
Source code and example input files included.

The user needs to provide a folder with structures (in .cif format) with a list of frameworks in the same folder.

See simulation.input file included for more information.

gRASPA translated to SYCL/C++

A version of gRASPA with GPU kernel functions translated to SYCL from CUDA.
For those who are curious about different devices.
It can be used for non-NVIDIA devices (Intel/AMD GPU cards), or even non-GPUs (parallel CPUs, FPGAs). It was reviewed around Dec. 2023.

Read more about SYCL: https://www.khronos.org/sycl/

gRASPA-FastOption-GCMC-NVT-Gibbs

The faster (simpler) version, with somewhat limited features.
SIMPLE, FAST, BETTER SCALING ON GPU.

• This code is able to perform:

1. Single component Grand Canonical Monte Carlo (GCMC)
2. Transition Matrix Monte Carlo (TMMC) in the GC ensemble (GC-TMMC)
3. Single component NVT-Gibbs Monte Carlo (tested with ethane vapor-liquid equilibrium)

• Difference from the main version:

1. This version only reports the total energy of the simulation
2. Faster speed and much better scaling when using Nvidia-MPS for single-component GCMC.
3. Please use the NVC_COMPILE in the zip folder attached to compile this code (mentioned later).

• NOTE:

1. For single-component GCMC, if you run the same simulation with the same setup and random seed, you should get the same result using this and the main version of gRASPA.
2. Please refer to the manuscript for the explanation of the fast option.

NOTE: download the first two in the assets below