High-Process Computing systems offer great potential in the field of environmental research, due to their ability to effectively deep-train artificial intelligence to aid in the accumulation of data. And yet, in the UK, environmental scientists’ access to such systems has been limited. But with funding from the Natural Environment Research Council (NERC), OCF has developed the Massive GPU Cluster for Earth Observation (MAGEO) to support scientists in their pursuit of combining deep learning and environmental research.
Artificial Intelligence has many applications, and through machine learning, it can be trained to become even more competent. Deep learning is a subset of machine learning that involves training AI to learn in a similar way to the human brain, noticing patterns and learning from mistakes, through the use of artificial neural networks. It is a longer process than machine learning but yields much more accurate results.
Such processes are ideal for scientists in fields such as environmental research, as deep learning algorithms can help them process a greater amount of data from the numerous sources from which they work. Scientists who sought to apply deep learning tools to their data, or deep learning experts looking to apply their techniques to Earth Observation data were at a disadvantage, prompting the Natural Environment Research Council to commission the creation of a specialist supercomputer.
OCF answered the call by developing the Massive GPU Cluster for Earth Observation, also known as MAGEO at the Plymouth Marine Laboratory (PML), a marine research organisation and registered charity.
MAGEO is built around five special edition NVIDIA DGX-1 MaxQ nodes that are more
energy-efficient, using 50 per cent of the power of a standard DGX-1 while still delivering 80 per cent of the capacity. Each DGX-1 contains 40,960 CUDA cores, 40 Intel Xeon 2.2 GHz cores and 512 GB system memory. The cluster has a dedicated 500TB of fast storage as well as the ability to access satellite data stored on the existing file system at PML with over 3 PB of storage.
MAGEO was designed with a diverse range of intended uses, from wildfire, oil spill and plastics detection, mangrove mapping and satellite and airborne data fusion. But it got a chance to prove its processing power before even being fully installed into PML’s data centre, as it was enlisted in the fight against COVID-19.
Professor Steve Widdicombe, the PML Director of Science said:
“This supercomputer will soon be hard at work processing satellite and air-borne data as part of crucial environmental research, but it’s great to see this powerful technology being put to such an important use in the meantime”.
Soon after being installed, MAGEO was used to develop a new method for the estimation of chlorophyll concentrations in the upper ocean. The AI training process for this would usually take 16 months, but with MAGEO’s help, it took only ten days.
Dr Daniel Clewley, Manager of NERC’s Earth Observation Data Acquisition and Analysis Service, said:
“AI is a tremendous step forward within the Earth Observation field. A staggering amount of data is produced by satellite every day and by using AI, researchers can tease out changes, trends and anomalies not possible with standard analysis. The new MAGEO HPC cluster provided and supported by OCF will enable researchers in the UK to fully utilise AI with Earth Observation data.”
Article By: Kim Nguyen
Latest case studies and news in the industry
Whatever your query or question, we would love to hear from you.
Please complete the form below and we will get back to you as soon as possible.
