The Wellcome Sanger Institute, renowned for its contributions to the Human Genome Project, is taking a leap forward in cancer research by incorporating NVIDIA's advanced computing solutions. This move aims to accelerate genomic analysis while simultaneously cutting down on energy usage and operational expenses.
The institute's Cancer, Ageing and Somatic Mutation (CASM) programme is at the forefront of this technological adoption. CASM processes tens of thousands of cancer genomes annually, providing critical insights into cancer formation and treatment efficacy. To manage this colossal workload, the Sanger Institute is exploring the use of NVIDIA DGX systems coupled with NVIDIA Parabricks, a genomics analysis software suite designed for accelerated computing.
The institute's adoption of NVIDIA technology has already shown promising results. In comparison to their current setup of 128 dual-socket CPU servers, a single NVIDIA DGX system has demonstrated remarkable improvements. Runtime has been reduced by 1.6 times, costs have been cut by 24 times, and energy consumption has been lowered by up to 42 times.
The Sanger Institute could potentially save $1 million and 1,000 megawatt-hours of energy each year by switching to NVIDIA GPUs for genomic analysis. To put this into perspective, the energy saved is equivalent to powering an average American home for a century.
The institute's NVIDIA-accelerated sequencing lab represents a new paradigm in scientific computing, functioning as an "AI factory" where raw data is transformed into valuable intelligence. This approach aligns with the concept of next-generation data centres designed to handle the most computationally intensive tasks with maximum efficiency.
In pursuit of further optimisations, the Sanger Institute is collaborating with Schneider Electric, a leader in energy management and automation. This partnership aims to minimise data center downtime and implement energy-efficient technologies such as uninterruptible power supplies and advanced cooling equipment.