Atlanta played host to the technological juggernaut that is Supercomputing 2024 (SC24) this year, bringing together the brightest minds in high-performance computing (HPC), artificial intelligence (AI) and scientific research. The conference didn’t just showcase the latest advancements—it set the stage for the future of computing.
From landmark announcements to jaw-dropping product showcases, this year’s event reminded us that the boundaries of innovation are always up for renegotiation. At the heart of it all, Boston Limited demonstrated its leadership in HPC, liquid cooling and AI, offering solutions as forward-thinking as the challenges they address.
Breaking Records, Redefining Limits
If there was one headline that stole the show, it was the unveiling of El Capitan as the new leader of the Top500 supercomputers list. Clocking a peak performance of 1.72 exaflops—1.72 quintillion calculations per second—it’s a machine built for the future. Housed at Lawrence Livermore National Laboratory, El Capitan isn’t just about raw power; it’s an emblem of what’s possible when you combine innovation with ambition. Its applications, from national security simulations to advanced climate modelling, are as awe-inspiring as the machine itself.
Not far behind in the buzz rankings was Cerebras Systems. Known for its bold approach to computing hardware, Cerebras announced a breakthrough in molecular dynamics: 1.1 million simulations per second, a feat made possible by its revolutionary wafer-scale architecture. For those wondering what the future of specialised computing might look like, Cerebras offered a glimpse—and it’s fast, efficient and game-changing.
AI Meets the Enterprise
AI was, unsurprisingly, everywhere at SC24. NVIDIA, always a crowd favourite, revealed updates to its Hopper architecture with the H200 NVL—a new solution designed specifically for enterprise-level AI in compact environments. While NVIDIA continues to dominate the AI accelerator market, this year felt like a message: AI isn’t just for research labs anymore. It’s coming to offices, factories and businesses that previously only dreamed of leveraging its power.
Another standout from NVIDIA was the introduction of cuNumeric, a toolkit that promises to simplify the scaling of Python programs to GPU clusters. For researchers and developers, it’s a bridge between simplicity and the scale needed to tackle the complexities of modern AI.
Sustainability also took centre stage, with NVIDIA showcasing its ALCHEMI NIM platform, dedicated to renewable energy materials research. This demonstrated how HPC and AI are being harnessed for planetary good.
Boston Limited: At the Heart of HPC and AI Innovation
Amidst the buzz and breakthrough announcements, Boston’s booth at SC24 was a hub of innovation, offering visitors the chance to engage directly with technologies shaping the future of high-performance computing and AI. As part of the SourceCode portfolio of companies, Boston leverages a global network of expertise to deliver cutting-edge solutions tailored to the most demanding workloads.
.jpg)
Boston’s portfolio, developed in collaboration with industry leaders like AMD, NVIDIA, Supermicro, Intel and Castrol, showcased its role as a trusted partner for organisations at the forefront of HPC and AI. By combining advanced hardware, sustainable technologies and unparalleled support, Boston continues to redefine the standards of innovation in the computing landscape.
AMD Instinct MI300X
A standout at Boston’s booth was the AMD Instinct MI300X accelerator, built for generative AI and large-scale model training. This advanced hardware enables groundbreaking research and business innovation, turning ambitious AI projects into reality. Visitors were encouraged to experience its capabilities firsthand by booking hands-on testing sessions at Boston Labs, offering a rare opportunity to explore its transformative potential in real-world environments.
Glacier Core Immersion Cooling Systems
Sustainability was a major focus at Supercomputing 2024, and Boston showcased its commitment with the Glacier Core immersion cooling systems. These solutions, leveraging advanced thermal management technology, drastically reduce energy consumption while maintaining top-tier performance. As the industry shifts toward greener practices, Glacier Core aligns perfectly with the growing demand for environmentally conscious high-performance computing solutions.
NVIDIA Grace Hopper GH200
Another highlight from Boston’s exhibit was the NVIDIA Grace Hopper GH200 Superchips, specifically designed for memory-intensive workloads. These chips empower organisations managing massive datasets, enabling efficient processing that turns complex ambitions into achievable realities. Perfect for AI and HPC applications, the GH200 demonstrates Boston’s dedication to delivering solutions that meet the evolving demands of modern enterprises.
.jpg)
Intel Gaudi 3 Processors
Intel’s Gaudi 3 AI processors demonstrated Boston’s versatility in delivering solutions tailored to deep learning applications. At Supercomputing 2024, Boston showcased the Gaudi 3 PCI-E card, which can be used in an 8-way GPU server from Supermicro, delivering exceptional performance for AI workloads. These processors bring efficiency and scalability to the table, enabling businesses to accelerate AI development without breaking the bank, while also supporting advanced configurations for intensive training and inference tasks.
.jpg)
Supermicro X13 Platforms
Supermicro’s X13 platforms feature robust support for 4th Gen Intel Xeon Scalable processors, delivering the performance and flexibility needed for cutting-edge HPC and AI workloads. At Supercomputing 2024, Boston also highlighted the Supermicro Xeon 6 server, a standout in the X13 range. This system is purpose-built for compute-intensive applications, offering powerful multi-node configurations for maximising performance and efficiency.
.jpg)
Castrol ON Liquid Cooling Fluid
Boston also had Castrol’s innovative liquid cooling technology on the booth, demonstrating how its cutting-edge Castrol ON fluids deliver superior thermal management. These dielectric fluids are engineered to maximise heat transfer efficiency while being non-conductive, ensuring safe and reliable cooling for high-performance computing systems. By significantly reducing the energy required for active cooling, Castrol’s solution contributes to sustainability goals, aligning with the industry’s shift toward greener, more efficient technologies.
.jpg)
You can find out more about how we’ve teamed up with Castrol ON and Submer here.
Ampere MegaDC Servers
Ampere’s MegaDC servers were a key feature of Boston’s Supercomputing 2024 display, with their scalability and efficiency for cloud-native and hyperscale applications. Complementing this was the introduction of Ampere’s upcoming Ampere One CPU, promising to set new energy-efficiency benchmarks for high-density environments. Boston also highlighted advanced processors, including the 9005 Turin AMD CPU, the NVIDIA H100 NVL GPUs for large language model training, the Gaudi 2 processors for deep learning tasks and the Untether AI accelerator for high-throughput, energy-efficient AI inference.
.jpg)
AMD Servers with ZutaCore DLC
On the Boston booth, we also had AMD’s servers featuring Zutacore’s Direct Liquid Cooling (DLC) solutions. These systems integrate AMD’s high-performance processors with advanced liquid cooling technology to achieve exceptional thermal efficiency. Zutacore DLC ensures consistent temperatures even during peak computational loads, making it an ideal choice for HPC and AI applications.
.jpg)
Boston Igloo AI+
The Boston Igloo AI+ brought cutting-edge storage solutions to the forefront at Supercomputing 2024. Tailored for AI and deep learning applications, the Igloo AI+ is designed to handle the data-intensive demands of modern enterprises. With its focus on high-performance storage and seamless scalability, it complements Boston’s extensive lineup of AI-optimised hardware, ensuring customers have the tools they need to unlock the full potential of their data-driven workloads.
.jpg)
The AI Factory Vision
Boston didn’t just showcase products; we shared a vision. As an NVIDIA SuperPOD Specialist, Boston is helping businesses realise the dream of building their own AI factories—end-to-end infrastructures designed to harness the full power of artificial intelligence. This isn’t about plugging in some GPUs and hoping for the best; it’s about architecting systems that deliver consistent, scalable results.
Boston’s SuperPOD offerings are complemented by a commitment to customer support, ensuring that organisations can deploy and maintain these AI factories with confidence. Our message is clear: Boston isn’t just a vendor; we're a partner for success.
Beyond the Booth: Trends and Takeaways
This year’s event was about the ideas driving HPC and AI forward. Quantum computing, for instance, is no longer a far-off dream but a growing part of the HPC ecosystem. Collaborations like Infineon and Quantinuum are bridging the gap between classical and quantum systems, opening doors to new computational possibilities.
Another major development was the introduction of Azure HBv5 virtual machines, which redefine what’s possible in terms of memory bandwidth for HPC workloads. For researchers and enterprises alike, these advances mean tackling problems faster and more efficiently than ever before.
The Road Ahead
The future of computing is one of collaboration, sustainability and relentless innovation and Boston’s presence at SC24 encapsulated this ethos, showcasing not just the best in technology but also the vision and expertise to help organisations make it their own.
For those who missed the event, Boston offers a wealth of resources to explore these cutting-edge solutions. Whether you’re looking to test the latest accelerators or build an AI factory from the ground up, the Boston team is ready to guide you every step of the way.
Reach out today and discover how Boston Limited can help you navigate the ever-changing world of HPC and AI.