The adoption of AI in businesses comes with collateral challenges related to computational power, bandwidth and memory demands, energy efficiency and sustainability, and AI talent gap.
As AI transforms industries, supercomputers are essential for breakthroughs in AI training, climate modeling, drug discovery, and more. Yet, the immense computational power required for these advancements brings challenges like high bandwidth, memory demands, and the growing urgency of energy efficiency.
Asia-Pacific enterprises are emerging as strong contenders in the global supercomputing race, frequently earning top ranks alongside the US and Europe. However, they face persistent hurdles, including inefficient data transfer, a limited talent pool, and escalating energy costs to optimize these complex systems.
Hewlett Packard Enterprise (HPE) has made significant strides with El Capitan, the world’s fastest supercomputer, delivering 1.742 exaflops of power. Joseph Yang, General Manager of HPC and AI, APAC and India, HPE, is therefore in a good position to address DigiconAsia’s questionssurrounding the critical challenges in scaling AI and high-performance computing while advancing sustainability goals.
What are the key factors driving supercomputing technology forward in Asia Pacific?
Joseph Yang (JY): Key factor driving supercomputing technology forward in Asia Pacific (APAC) include the rapid advancements in artificial intelligence (AI), which require immense computational power for processing large-scale data sets and solving complex global challenges such as climate change, health, and food security. Emerging technologies such as quantum computing are also shaping the next generation of supercomputers, offering unprecedented speeds and efficiency.
Additionally, APAC nations, notably Singapore, are heavily investing in supercomputing to advance national research initiatives and drive industry transformation. Recently, National Supercomputing Centre (NSCC) in Singapore has signed an MoU with CSC Finland, which operates LUMI, one of the EU’s most powerful supercomputers built by HPE, to advance HPC adoption, share best practices, and strengthen capabilities in AI, HPC-Quantum, digital twin, and green data center technologies. The region’s competitive position alongside the US and Europe underscores its commitment to computational science innovation.
The introduction of these advanced systems strengthens the region’s capability to lead in cutting-edge research and address increasingly complex technological and environmental challenges.
How do investments in high-performance computing deliver long-term value for businesses despite cost and operational challenges?
JY: Investments in high-performance computing (HPC) deliver long-term value for businesses by enabling them to tackle complex challenges and unlock insights that traditional computing cannot handle. While HPC systems can be costly and operationally demanding, solutions like clustered smaller computers or HPC as a service offer cost-effective alternatives. Businesses can scale resources as needed, paying only for what they use.
HPC enables advancements in key areas such as fraud detection, medical record management, and weather prediction, allowing companies to process large data sets quickly and accurately. This results in increased efficiency, innovation, and competitive advantage. Additionally, with the rise of accelerated computing and AI, HPC enhances capabilities in modeling, simulation, and data analytics, driving digital transformation across industries.
By adopting these technologies, businesses can stay ahead in the ever-evolving digital landscape, ensuring long-term growth despite initial investments and operational challenges.
How is El Capitan addressing the data transfer and energy efficiency challenges in supercomputing?
JY: El Capitan addresses data transfer and energy efficiency challenges in supercomputing through innovative technologies. It uses HPE’s Cray EX supercomputing platform powered by AMD’s MI300A APUs, which integrate CPU and GPU cores with high-bandwidth memory. The system is supported by HPE Slingshot, a high-speed Ethernet fabric, allowing fast and efficient data transfer across more than 11,000 nodes.
In terms of energy efficiency, El Capitan leverages the industry’s first 100% fanless direct liquid cooling system. This solution reduces cooling power by 37% per server blade, cutting utility costs, and lowering carbon emissions. The system’s high-density design also enables greater server cabinet density, consuming less floor space and supporting sustainable, high-performance computing.
