Pakistan Supercomputing Centre

Pakistan Supercomputer Centre is a high-performance computing facility associated with PakHPC and hosted at the Centre for AI & Big Data (CAID), Namal University, Mianwali, Punjab, Pakistan.¹² The facility supports research and development in artificial intelligence, big data, distributed computing, modelling and simulation, cloud computing, and real-time industrial applications.¹

The system has been described in IET Conference Proceedings as an open-source, scalable and heterogeneous supercomputing platform designed for high-performance computing applications.³

The technical background of Pakistan Supercomputer Centre (PakSupercomputer) is linked to earlier work on heterogeneous supercomputer architecture in Pakistan. In January 2016, The Nation reported that a team of engineers at the Faculty of Computing and Applied Science, Riphah International University, had developed a supercomputer architecture. The report named Dr. Tassadaq Hussain, Engr. Tariq Ali and Abdur Rehman among the engineers involved in the work.⁴

The 2016 report stated that the system supported CUDA, MPI/LAM, OpenMP, OpenCL and OpenACC programming models and was intended for applications including numerical computing, big data, data mining, bioinformatics, genomics, business intelligence, analytics, climate, weather and ocean-related problems.⁴

A later project titled Development Of A Scalable Heterogeneous Supercomputer was funded through the Technology Development Fund (TDF) of the Higher Education Commission of Pakistan. According to UCERD, the project received HEC Technology Development Fund support of Rs. 16 million, with Dr. Tassadaq Hussain as principal investigator and UCERD Pvt. Ltd. as industrial partner.⁵

PakSupercomputer is hosted at the Centre for AI & Big Data at Namal University in Mianwali, Punjab, Pakistan.¹ Namal University identifies CAID as a research and development centre focused on artificial intelligence and big data. Dr. Tassadaq Hussain is listed by Namal University as Professor and Director CAID.²

CAID describes its research groups as covering supercomputing and cloud computing, real-time industrial applications, secure digital system design, and AI and big data applications.¹

CAID publicly lists the Namal HPC infrastructure as a 1.6 petaFLOPS high-performance computing system. The listed hardware includes 20 nodes, 1600 processor cores, 5 TB of main memory, 40 TB of SSD storage, 10 Gigabit Ethernet networking, a low-latency switch, and 40 Nvidia RTX 4070 Ti GPUs for distributed acceleration.¹

CAID states that the cluster handles computational simulation and compilation tasks with high-speed data transfer and low-latency node-to-node communication.¹

The IET Conference Proceedings paper describes PakSupercomputer as a heterogeneous supercomputing platform. The system is presented as combining heterogeneous processing resources with an open-source distributed system stack and an elastic parallel programming model.³

The UCERD project page states that the scalable heterogeneous supercomputer model supports CUDA, MPI/LAM, OpenMP, OpenCL and OpenACC programming models. It also describes the use of an elastic parallel programming model intended to execute application tasks on a heterogeneous parallel computing platform through a user-friendly web-based interface.⁵

The facility supports work in high-performance computing, artificial intelligence, cloud computing, modelling and simulation, secure digital system design, and real-time industrial applications.¹ CAID lists projects in indigenous supercomputing, VR tele-rehabilitation and smart rice sorting, with partners including Barcelona Supercomputing Center, Pakistan Supercomputing Center, New Zealand Chiropractic Center, Riphah Rehab Center and Alkaram Rice Engineering.¹

The UCERD project page identifies possible application areas including numerical computing, big data and data mining, bioinformatics and genomics, business intelligence and analytics, climate modelling, weather and ocean-related problems.⁵

The Centre launched Pakistan's National Centre of Excellence in Supercomputing in collaboration with Ghulam Ishaq Khan Institute of Engineering Sciences and Technology and Pak-Austria Fachhochschule: Institute of Applied Sciences and Technology.⁶ The initiative aims to provide HPC infrastructure, standardized processes, and specialized talent to address national challenges in health, agriculture, and industry, while enabling shared HPC access, indigenous AI development, and real-world industry deployments across Pakistan.⁷

The Centre for AI & Big Data conducts training and outreach activities related to artificial intelligence, big data and high-performance computing.¹ Namal University has also hosted training activity under the title Supercomputing for AI and Big Data Applications.⁸

PakistanSupercomputing is the public-facing initiative and web presence associated with PakSupercomputer. It uses the PakHPC domain and presents work in supercomputing, artificial intelligence, distributed computing, cloud computing, edge computing, bare-metal systems and hardware-software co-design.⁹

PakistanSupercomputing is a registered trademark with the Intellectual Property Organization of Pakistan. The name and logo were advertised in The Trade Marks Journal, Journal No. 858, page 869 dated 1 July 2022, under application number 621610, and were later registered under the issued trademark certificate.¹⁰