Understand this story
Ferveret announced that researchers at MIT's Extreme Environment Materials Laboratory are using its Adaptive Phase Cooling high-performance computing platform to conduct two studies: AI analysis of coherent X-ray measurements of nuclear materials and machine-learning interatomic potentials. Ferveret says the solution can improve server-level compute by up to 15% compared with leading direct-to-chip cooling, and by up to 35% in usable compute when facility-level improvements are included. Of this, 15% comes from a per-kilowatt TFLOPs efficiency improvement measured by a UCLA laboratory using NVIDIA H200 GPUs; the 35% is extrapolated from a hypothetical scenario in which PUE drops from 1.20 to 1.03. There are currently no further developments.
AI-generated from public reports
Latest development · Report
MIT Researchers Use Ferveret Cooling Platform to Advance AI Research on Nuclear MaterialsDevelopments
1 confirmed developments
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Oct 23
Unite.AI
MIT Researchers Use Ferveret Cooling Platform to Advance AI Research on Nuclear Materials
MIT's Materials Extreme Environment Laboratory is using Ferveret's Adaptive Phase Cooling high-performance computing platform to conduct two research projects: AI analysis of coherent X-ray measurements of nuclear materials and machine-learning interatomic potentials. Ferveret says its solution can improve server-level compute by up to 15% compared with leading direct-to-chip cooling, and by up to 35% in usable compute when facility-level improvements are included; the 15% comes from a per-kilowatt TFLOPs efficiency gain measured by a UCLA laboratory using NVIDIA H200 GPUs, while the 35% is extrapolated from a hypothetical scenario in which PUE drops from 1.20 to 1.03.