Accurately and promptly prescribing treatments is crucial in healthcare, but it is challenging due to the complexity of the human body and disease progression. Doctors often rely on informed judgment and adjust treatments based on patient responses. Developing and delivering new medications is also often complex and slow, particularly when researchers explore new mechanisms or biological pathways with uncertain disease links.
To tackle these challenges, healthcare organisations are digitalising their operations and using digital tools like wearables to collect more data. However, data alone is inert; it needs processing power to be converted into actionable insights. This is where high-performance computing (HPC) can make a difference.
HPC can process massive, multi-dimensional data sets to solve complex problems at extremely high speeds. It runs multiple tasks simultaneously on clusters of computer servers and processors. According to IBM, HPC systems typically run at speeds more than one million times quicker than the fastest commodity desktop, laptop or server systems.
“HPC is a powerful tool for medical research and healthcare innovation, helping the industry deliver quality care to more patients and help doctors make better decisions faster. It is revolutionising the healthcare landscape by creating new ways to diagnose and treat diseases, driving significant improvements in image processing, sparking drug discoveries, powering genetic-level research and empowering healthcare professionals with greater precision,” says Peter Chambers, managing director for Asia Pacific and country manager for Australia at AMD.
One of the key applications of HPC in healthcare today is data analysis, says Miles Upton, regional general manager for Asia at Cambridge Consultants. By harnessing extensive data from electronic patient records, connected devices and clinical studies, healthcare providers can gain deeper insights that could improve patient outcomes through personalised care.
HPC is also used beyond healthcare delivery, such as for drug discovery and clinical research. “When combined with artificial intelligence (AI), HPC can drastically reduce the time and cost associated with identifying and designing new drug candidates, ultimately speeding up the process of bringing new drugs to market. Another innovative application of HPC is in data augmentation, where synthetic data generation can extend the relevance and scope of clinical studies. This approach reduces the burden on actual clinical trials and improves the coverage of the intended patient population,” adds Upton.
See also: Alibaba anoints new chief in revamp of stalling commerce arm
Supporting genomics
Many diseases, including cancers, are caused by changes in our genes. Genomics, the study of genes, enables more precise prediction, diagnosis and treatment of diseases. Viruses also have genomes, so scientists can use genome sequencing — determining the complete DNA or RNA sequence of a virus’s genome — to track how viruses like Covid-19 spread and mutate, aiding vaccine development.
The challenge in genomics lies in managing the vast amount of data. The human genome consists of 3 billion DNA bases of four different types, represented by four “letters” in the DNA “alphabet.” In contrast, the coronavirus has an RNA genome of 30,000 letters, about 100,000 times smaller than the human genome.
See also: Break up Google? What’s at stake in antitrust action
Scientists must analyse data from genome sequencing alongside external data from environmental sensors and wearables to understand how a virus impacts humans and to develop potential cures.
“Predicting virus evolution has always been a challenge. Through genome sequencing, researchers have created maps of gene variations related to health, disease and drug response. This data deluge, coupled with the need to integrate diverse data types, poses a major challenge for research institutions lacking robust computational infrastructure for analysis. [They need to be supported by] HPC and advanced data management solutions,” says Sinisa Nikolic, director of Asia Pacific for HPC-AI and cloud service provider at Lenovo.
Recognising this, Lenovo offers a Genomics Optimization and Scalability Tool (GOAST), which uses HPC and AI to help researchers quickly map the genomes of a cohort of people instead of analysing a single genome. HPC supports high-throughput volumes to accelerate the speed of analysis, while AI helps make sense of the differences between genomes.
“GOAST enables scientists to process more samples daily, decreasing time to discovery from days and hours to minutes. Scientists can also analyse extensive datasets as GOAST leverages powerful parallel processors. This is vital for understanding complex cellular processes, improving disease prevention and tracking biological evolution,” says Nikolic.
“For example, a supercomputing centre in Barcelona used GOAST to shorten the time to identify variants or mutations in DNA samples (also known as germline variant calling) from 30 hours to 45 minutes and enhance its throughput capacity. By improving lab productivity and enabling the simultaneous processing of more genomes, GOAST supports more efficient and comprehensive research, advancing precision medicine and life-saving discoveries.”
There are also other HPC-based solutions for genomics. “Illumina’s NovaSeq X sequencing systems, for instance, leverage AMD EPYC high-performance CPUs and Alveo U250 accelerator cards to provide massive throughput and productivity gains that enable sequencing of up to tens of thousands of genomes per year. It can, therefore, fuel groundbreaking advancements in oncology, reproductive health, genetic disease, microbiology, agriculture, forensic science and beyond,” adds AMD’s Chambers.
Accelerating other medical research
To stay ahead of the latest tech trends, click here for DigitalEdge Section
HPC can also advance medical studies beyond genomics. For example, the Research Institute for Nanodevice and Bio Systems at Hiroshima University in Japan uses AMD’s HPC solutions to transform the traditionally human-led task of diagnosing abnormalities during a colonoscopy.
“In the past, diagnoses were dependent on the experience and expertise of the physician. With a system leveraging AMD’s Alveo U250 accelerator cards and Zynq UltraScale+ MPSoC, physicians at the institute can use AI to quantify the progress of a lesion based on colonoscopy image data. This empowers them to make more accurate diagnoses faster and deliver personalised care and treatment to patients quickly,” says Chambers. In HPC systems, MPSoCs (multiprocessor systems-on-chip) handle the overall system management, while accelerator cards focus on specific computational tasks. This combination allows for efficient utilisation of resources and improved performance of HPC systems.
Another use case for HPC is motion-based drug design or molecular dynamics simulation, which allows drug designers to see the entire effective process of drug molecules more directly.
Shanghai-based Snowlake Technology, for instance, wanted to build a cost- and power-efficient HPC solution that could produce enough computing power for molecular dynamics calculations. It, therefore, developed Yaddle MD, a dedicated solution for molecular dynamics computing on AMD Alveo U200 accelerator cards.
AMD Alveo U200 accelerator cards are designed to meet the constantly changing needs of the modern data centre, providing up to 90 times performance increase over CPUs for common workloads, including machine learning inference, video transcoding and database search and analytics. As such, Yaddle MD achieves nearly 50 times the performance of CPUs and twice that of the industry’s highest-performing GPUs, at a much lower power consumption.
Peter Chambers, managing director for Asia Pacific and country manager for Australia, AMD
Inhibited by cost
Cost is a major barrier to the widespread adoption of HPC in healthcare, but Nikolic from Lenovo believes that the long-term benefits will outweigh the expense. He adds: “For healthcare organisations starting with HPC infrastructure, the initial costs may be high due to hardware and set-up expenses. However, HPC is highly cost-effective for preventive healthcare in the long run, particularly when combined with AI and other advanced technologies. HPC can drastically reduce the time and power needed for genomic sequencing and complex data analyses.
“Integrating AI with HPC enhances these benefits further. AI-assisted diagnostics, predictive analytics, 3D modelling and remote reading leverage HPC’s power to improve diagnostic accuracy and outcomes. This combination minimises errors, enables earlier interventions and supports more precise therapies, benefiting both doctors and patients.”
Healthcare organisations can also turn to HPC-as-a-service (HPCaaS) to access HPC without substantial upfront investments and gain scalability. “Since HPCaaS provides a flexible, pay-as-you-go model, healthcare organisations can scale their resources according to current needs. AI integration with HPC optimises this scalability by adapting to evolving data and research demands. From a macro perspective, organisations that use everything-as-a-service offerings gain access to not only HPC solutions but also the ability to manage their entire infrastructure more seamlessly,” says Nikolic.
Lenovo’s GOAST, he says, costs up to 50% less than boutique solutions relying on GPUs or FPGAs (Field Programmable Gate Arrays to power custom, low-latency applications) without additional licensing fees.
A single GOAST server can replace up to 50 standard nodes. Lenovo’s HPCaaS offers the performance and flexibility to manage capital expenditure and operating expenses effectively, supporting various applications from genomics to pharmaceuticals. When combined with AI, HPC extends the diagnostic strength of major hospitals to front-line health centres, improving accuracy, reducing errors and enhancing patient outcomes.
Sinisa Nikolic, director of Asia Pacific for HPC-AI and cloud service provider, Lenovo
Succeeding with HPC in healthcare
Due to its high computational demands, HPC can only provide its benefits if a robust IT infrastructure supports it.
“This further reinforces the need to consider infrastructure-as-a-service for small and large healthcare organisations to optimise their investments and access advanced technology to improve healthcare. Having a poor IT infrastructure while running HPC workloads is akin to having the best engine fitted in a car body that is 20 years old,” says Nikolic.
Implementing data security protocols to protect sensitive patient information is essential for regulatory compliance, especially given the healthcare industry’s vulnerability to cyberattacks. According to cybersecurity company Sophos, 67% of healthcare organisations experienced ransomware attacks in 2024, with 74% of these attacks resulting in data encryption.
Ethical considerations are another important aspect that healthcare organisations should address. “Fairness and the responsible use of genetic data have always been a pertinent topic of interest, especially when privacy was a concern in developing the vaccine [using HPC] during the Covid-19 pandemic,” says Nikolic.
Healthcare organisations, he adds, need to ensure their staff has the expertise required to maintain and operate HPC systems and infrastructure. “Going back to my car analogy, you can have a good car and engine, but you also need a good driver to maximise its potential.”
Enabling precision medicine
The future of healthcare lies in precision medicine, where treatments and prevention strategies are customised to each patient’s unique genetic, environmental and lifestyle factors. Achieving this requires a blend of expertise and advanced technologies, including HPC.
HPC advancements will immensely benefit genome sequencing. The ability to process large genomic datasets rapidly enables faster and more accurate DNA sequencing, which is crucial for understanding genetic diseases, identifying mutations and developing personalised treatment plans. Routine genetic screening, powered by HPC, could soon allow for the diagnosis of diseases before symptoms even emerge, transforming preventive healthcare.
Miles Upton, regional general manager for Asia, Cambridge Consultants
He also highlights the potential of using HPC for digital twins in healthcare. “Digital models [of patients’ bodies] allow healthcare organisations to simulate changes in therapy, lifestyle and environmental factors before they occur, offering a powerful tool to optimise treatments and improve medical outcomes. This proactive approach can revolutionise personalised healthcare by tailoring interventions based on simulated results.”
AMD’s Chambers adds: “The quest to understand the human body never rests. Many aspects of our complex and dynamic molecular networks remain mysterious and HPC is changing that by enabling genomic approaches and sophisticated analytics that help explain the human body’s awe-inspiring sophistication. By doing so, scientists discover novel insights that help fight disease and improve lives.”
