Performance Impact of Containerized METADOCK 2 on Heterogeneous Platforms
Antonio Jesús Banegas-Luna, Baldomero Imbernón Tudela, Carlos Martínez-Cortés, José María Cecilia, Horacio Pérez-Sánchez · 2025
Read Performance Impact of Containerized METADOCK 2 on Heterogeneous Platforms by Antonio Jesús Banegas-Luna, Baldomero Imbernón Tudela, Carlos Martínez-Cortés, José María Cecilia, Horacio Pérez-Sánchez (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2506-06450.txt.
Introduction
Antonio Jesús Banegas-Luna and four co-authors test whether packaging a drug screening program into software containers slows it down, in this 2025 paper. Virtual screening docks millions of candidate molecules against protein targets and eats computing clusters for breakfast. Containers promise portable, reproducible runs, but scientists fear a performance tax.
The docking tool was run under three container systems, Docker, Singularity, and Apptainer, across mixed processor and graphics configurations. The overhead came out negligible, under one percent in every setup. The tool also handled large molecular complexes that strain a popular commercial rival. The authors conclude that containerized deployment is safe for high throughput screening: portability and reproducibility arrive essentially free, which eases sharing pipelines across the varied machines of modern computing centers.
Catalog record
| CREATOR | Antonio Jesús Banegas-Luna, Baldomero Imbernón Tudela, Carlos Martínez-Cortés, José María Cecilia, Horacio Pérez-Sánchez |
|---|---|
| YEAR | 2025 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 34 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2506-06450
Browse the full library · About the Tower of Babel
Full text
The opening of the work, from the library's hosted text:
Performance Impact of Containerized METADOCK 2 on Heterogeneous Platforms Antonio Jesús Banegas-Luna, Baldomero Imbernón Tudela, Carlos Martínez-Cortés, José María Cecilia, Horacio Pérez-Sánchez Abstract Virtual screening (VS) is a computationally intensive process crucial for drug discovery, often requiring significant resources to analyze large chemical libraries and predict ligand-protein interactions. This study evaluates the performance impact of containerization on M E T A D O C K 2 METADOCK\ 2 , a high-throughput docking software when deployed on heterogeneous high-performance computing (HPC) platforms. By testing three containerization technologies—Docker, Singularity, and Apptainer—across varying CPU and GPU configurations, the experiments reveal that containerization introduces negligible performance overhead, with deviations below 1%. Moreover, M E T A D O C K 2 METADOCK\ 2 demonstrated the capability to efficiently process large molecular complexes, surpassing the limitations of commercial tools such as AutoDock Vina. The results underscore the advantages of container-based deployment for ensuring portability, reproducibility, and scalability in scientific computing. This study concludes that containerized M E T A D O C K 2 METADOCK\ 2 is a robust and efficient solution for VS tasks on heterogeneous HPC platforms. keywords METADOCK, Containers, Molecular Docking, GPU, HPC, Metaheuristic † † equal-contributors: These authors contributed equally to this work. † † equal-contributors: These authors contributed equally to this work. † † equal-contributors: These authors contributed equally to this work. † † equal-contributors: These authors contributed equally to this work. † † equal-contributors: These authors contributed equally to this work. 1 Introduction Containers have revolutionized software deployment and execution by providing a lightweight, portable, and reproducible environment for applications Koskinen et al. 2019 . Unlike traditional virtualization, containers encapsulate applications and their dependencies in a single, isolated unit, allowing them to run consistently across different computing environments Mouat 2015 . This makes containers particularly advantageous for scientific computing, where reproducibility and scalability are critical Moreau et al. 2023 . One of the key benefits of containers is their ability to ensure compatibility across heterogeneous systems Hu et al. 2020 . Researchers and developers can package an application with all its dependencies, eliminating conflicts arising from differences in operating systems or libraries. This portability reduces the overhead of configuring software for different platforms, enabling seamless transitions between local development machines, cloud infrastructures, and high-performance computing (HPC) clusters Keller Tesser and Borin 2023 . Another advantage of containers is their efficiency. By leveraging the underlying host operating system rather than virtualizing an entire machine, containers achieve near-native performance. This is particularly valuable in computationally intensive fields like molecular docking Medeiros et al. 2023 ; Huff et al. 2023 , artificial intelligence Egbuna 2024 ; Chiang 2023 ; Theodoropoulos et al. 2023 , and data analysis Renton et al. 2024 ; Ru et al. 2023 ; Weisbart et al. 2023 , where even minor performance losses can have significant impacts on processing large datasets. Containers also enhance collaboration and sharing. Tools like Docker Hub and Singularity registries allow researchers to distribute pre-configured environments, ensuring that others can replicate their work without rebuilding complex setups. This fosters transparency and accelerates progress in multidisciplinary projects. In addition, containers are well-suited for scaling applications. Orchestration tools such as Kubernetes enable the deployment of thousands of container instances to handle dynamic workloads. This flexibility has made containers indispensable for modern cloud computing and big data pipelines Wang et al. 2023 ; Kumar et al. 2024 . Virtual screening (VS), which involves the computational screening of large libraries of compounds to identify potential drug candidates, can be time-consuming for several reasons. First, VS often consists of analyzing millions of chemical compounds, each evaluated against specific criteria, such as its fit to a target protein’s binding site or its predicted biological activity Cao et al. 2022 ; Walker and Clardy 2021 . Handling these vast datasets requires significant computational resources and time Crunkhorn 2022 . Secondly, the algorithms used in VS can be computationally intensive. Techniques such as molecular docking, molecular dynamics simulations, and pharmacophore modeling involve complex mathematical calculations and iterative simulations. These algorithms aim to accurately predict the interactions between molecules and biological targets, adding to the computational burden Yan et al. 2013 ; Stanzione et al. 2021 ; Brooks et al. 2021 . Finally, assessing each compound’s potential requires scoring functions to estimate its binding affinity or biological activity. These scoring functions often rely on energy calculations that account for multiple molecular interactions, making them computationally expensive, especially for large-scale screenings Murugan et al. 2022 . To address these challenges, HPC techniques, such as parallel computing Fan et al. 2021 , self-organizing maps Jayaraj et al. 2022 , and heterogeneous supercomputing systems Liu et al. 2023 , have become indispensable. However, modern chemical databases’ sheer scale and complexity necessitate a more robust solution. HPC clusters provide such a solution by enabling the distribution of massive datasets into smaller, manageable blocks that can be processed simultaneously across multiple computing nodes. This parallelism significantly reduces processing time while maintaining accuracy, making it essential for the success of large-scale VS projects. These advancements highlight the need for computational resources and the critical role of efficient algorithms and infrastructure in meeting the growing demands of drug discovery. Despite their many advantages Banchelli et al. 2020 , HPC clusters still need to deal with heterogeneity, which introduces some problems such as performance variability, data movement overhead, and portability Hutson et al. 2019 ; Khallouli and Huang 2022 . Ensuring portability across heterogeneous architectures is a significant challenge. Applications may need to be redesigned or optimized for each target architecture, limiting their portability and increasing development time and costs. Moreover, heterogeneous clusters may lack a unified software ecosystem, making it challenging to install and maintain software across different architectures. Compatibility issues and dependency management can hinder the deployment of applications and tools. Consequently, the use of containers is strongly recommended to overcome such issues. Containers offer a flexible and efficient approach to managing heterogeneity in HPC clusters, providing portability, isolation, and resource management capabilities that are essential for deploying and scaling applications across diverse hardware architectures. Among the many container technologies available, Docker Docker 2020 has become the most popular to deploy software applications or services in the last few years. Docker is widely used as a lightweight virtualization solution in several contexts Xie et al. 2020 ; Sharma et al. 2020 ; Kithulwatta et al. 2021 ; Kim et al. 2022 . However, there are some limitations to its scientific use. On the one hand, it may require some orchestration mechanism, which requires other tools such as Kubernetes Luksa 2017 and Docker Swarm Marathe et al. 2019 . On the other hand, Docker requires root privileges to build and manage containers, which can be a barrier for some scientific computing use cases Sultan et al. 2019 . To overcome these issues, Singularity and Apptainer containers have been developed to address specific challenges. Singularity and Apptainer are designed for scientific and HPC use cases where users may not have root access to the host system. It allows users to run containers themselves without requiring root privileges.