NAMD stands for Nanoscale Atomistic Molecular Dynamics, which is a program for simulating atomic-level molecular dynamics at the nanoscale.

The NAMD Benchmark dataset is a set of performance benchmark inputs and profiles designed specifically for high-performance computing (HPC) environments. It is used to evaluate the performance of the molecular dynamics simulation software NAMD under different hardware platforms, parallel modes (MPI, GPU, Charm++), and compilation optimization settings.

Unlike scientific experimental data or simulation results, this dataset contains standardized benchmark input/configuration packages used to measure the computational performance (speed, scalability, and efficiency) of a system, rather than simulation outputs for scientific analysis.
This benchmark is maintained by the NAMD official team—the Theoretical and Computational Biophysics Group and the Parallel Programming Laboratory at the University of Illinois Urbana-Champaign.

The relevant paper results areScalable molecular dynamics with NAMDThe dataset, titled "...", was released by Phillips et al. in 2005. Version of this dataset is "...".NAMD 3.0 beta 6 GPU-resident benchmarking results".

Dataset structure

This dataset typically contains:

• .namd is the main input file (controls simulation parameters).
• .psf and .pdb structure files (system topology and coordinates)
• .prm / .top force field parameter files (CHARMM format)
• Run the script and README

Dataset Contents

This dataset contains several representative molecular systems for performance testing at different scales and in parallel modes.

• APOA1
• System: APOA1 protein + solvent system
• Number of atoms: Approximately 92,000
• Features: Medium-sized benchmark system, often used for mixed-mode testing of CPU and GPU.
• STMV
• System: Satellite Tobacco Mosaic Virus
• Number of atoms: Approximately 1,066,628
• Features: NAMD's most commonly used performance standard system, used for large-scale parallel performance testing.
• STMV_21M / STMV_224M
• System: An extended copy of STMV
• Number of atoms: Approximately 21 million / 224 million
• Features: Used to test the scalability of massively parallel computing.
• F1-ATPase
• System: F1-ATPase enzyme kinetic simulation system
• Features: Complex protein system used to evaluate the efficiency of energy computation and constraint algorithms.

Instructions

• These benchmark files are used for performance testing and verification and do not contain scientific analysis results.
• It can run in single-node (CPU/GPU) or multi-node (MPI) environments to compare the efficiency of different hardware or versions.
• Users can test the impact of different timing lengths, cutoff radii, or PME settings on performance by modifying input parameters.
• The log results provided on the official page can be used as a benchmark to verify the correctness of the system configuration.