SMC Achieves Breakthrough Energy Efficiency in MLPerf™ Training with Single-Phase Immersion Cooling

Press Release | SMC Achieves Breakthrough Energy Efficiency in MLPerf™ Training with Single-Phase Immersion Cooling

Singapore, 13 June – Sustainable Metal Cloud (SMC), a leading AI GPU cloud platform powered by Firmus Technologies’ innovative single-phase immersion cooling technology, today announced groundbreaking results in the latest MLPerf™ training benchmarks.

As a recent member of ML Commons, SMC participated in the GPT-3 175B, 512 H100 Tensor Core GPUs submission, achieving world-class performance while demonstrating a significant reduction in energy consumption. When compared to GPU clouds hosted in traditional air-cooled data centers, SMC’s platform offers up to 50% energy savings – lowering the cost to use and carbon impact of AI development.

“MLPerf’s benchmarks exist to make machine learning better. It’s fantastic to see Sustainable Metal Cloud (SMC), one of our newest members, submit to MLPerf Training with our first ever power measurements,” said David Kanter, MLCommons Executive Director.

“We are thrilled to be a part of MLCommons and contribute to advancements in energy-efficient AI training,” said Edward Pretty, Chairman of SMC. “These results, verified by MLCommons members, validate the transformative power of our Sustainable AI Factories in reducing the environmental impact of large-scale AI. As the demand for AI grows, addressing resource consumption is critical. MLCommons has provided us the platform to validate our technology which now presents as a viable solution for governments and industry to host the next generation of AI compute, whilst maintaining world-leading performance, sustainably, using less power, water, and space.”

SMC’s Sustainable AI Factory Advantage

SMC’s unique single-phase immersion cooling technology, a cornerstone of its Sustainable AI Factories, plays a crucial role in achieving these exceptional results. This innovative design removes heat directly from the processors through submerging the servers directly into a liquid tank, eliminating the need for energy-intensive fans and chilled water-based air-conditioning systems typically found in traditional data centres. This leads to a 30% energy saving at the server level, with a further 20% energy saving from retrofitting the Immersion platform directly into traditional air-cooled data centre floor space.

David Kanter further added, “SMC’s release establishes a baseline for best practice power consumption. The MLPerf benchmarks help buyers understand how systems perform on relevant workloads. The addition of the new power consumption benchmark gives our members, buyers, and the entire AI community a new way to rate energy efficiency and environmental impact. As a community it’s important that we can measure things, so we can improve them. I hope SMC’s initial results will help drive transparency around the power consumption of AI training. This is an example of why MLCommons exists – to bring the best of industry together and have new, scaling tech platforms benchmarked against the world’s largest infrastructure providers.”

Further Optimisations Through Partnership

Furthermore, SMC is one of the world’s first clouds to enable VBOOST. By adopting an enhanced VBOOST-enabled software stack post-submission, it achieved further energy reductions to 451kWh and a 7% performance improvement, positioning the company’s customer cloud training performance environment just 6% below NVIDIA’s flagship Eos AI supercomputer1.

Democratising Sustainable AI

Sustainable Metal Cloud’s mission goes beyond achieving industry-leading efficiency. The company was founded upon the principle of democratizing access to sustainable AI. Sustainability comes in two forms: allowing everyone to build with AI while minimizing their environmental impact, yet doing so whilst delivering cost savings for the ecosystem.

Collaboration as a Driver of Innovation

SMC actively participates in MLCommons to not only showcase its technology but also to contribute to the development of industry-wide standards for energy-efficient AI training. The company believes transparency around energy consumption is essential for fostering long-term environmental accountability within the AI industry and works closely with leading technology partners like Dell Technologies to optimize its platform and explore new solutions that push the boundaries of energy efficiency.

Looking Forward: Sustainable AI for the Future

Sustainable Metal Cloud remains focused on supporting the development and deployment of cutting-edge AI models while promoting the widespread adoption of energy-efficient AI technologies. The company is actively investing in research to ensure its platform continues to deliver best-in-class performance for next-generation AI workloads while minimizing environmental impact.

About Sustainable Metal Cloud

Sustainable Metal Cloud (SMC) is a leading AI GPU cloud platform developed by Singapore-based Firmus Technologies. Leveraging its proprietary single-phase immersion cooling technology, SMC’s Sustainable AI Factories deliver benchmarked world-class performance with up to 50% total energy savings. SMC offers a globally expanding network of scaled GPU clusters and infrastructure, empowering customers to train AI models with unmatched efficiency and cost savings.

Media Contact:
Theodore Woon, Director

Pinpoint PR Pte. Ltd.

📞+65 9155 3507

Check out their website
🌐 theodore@pinpointpr.sg