The University of Strathclyde’s Continuous Manufacturing and Advanced Crystallisation (CMAC) medicines manufacturing research center has installed a new supercomputer dedicated to AI at its campus in Glasgow, Scotland.
Comprised of Nvidia DGX H100s, the university said the supercomputer is the first of its kind in a UK-based university and will enable CMAC to use AI to develop new medicines more quickly and affordably while cutting emissions.
The system will enable CMAC to use autonomous robotic platforms driven by hybrid AI and physics-based models, and according to CMAC will “dramatically enhance computational capabilities.”
CMAC has not shared the computational power capability of the new supercomputer, nor provided a name. Each H100 DGX features eight H100 GPUs, and provide provides 32 petaflops of compute performance at FP8 precision.
Funding for the supercomputer was provided by the UK Research Partnership Investment Fund (RPIF) Net Zero Pilot. The university was awarded £2.5 million ($3.2m) from the government to create a sustainable digitalized “lab of the future” and promised to reduce greenhouse gas emissions by over 50 percent in the labs.
Professor Blair Johnston, associate director of CMAC data management and infrastructure, said: “From novel ChatGPT-like large language models for medicines manufacturing to real-time imaging applications deployed at the edge on our advanced manufacturing process lines, Nvidia DGX H100 will allow us to advance research challenges previously beyond accessible computational capabilities. Working collaboratively with Nvidia's experts and pharma partners we are in a strong position to further accelerate the impact of CMAC’s portfolio of industrial digital technology research on medicines development in the coming years.”
The supercomputer was deployed with the assistance of Scan, an IT and infrastructure solutions provider.
“Scan feels privileged to be working with Nvidia and the University of Strathclyde’s CMAC facility to deliver and install the first UK-based DGX H100 solution,” said Elan Raja, CEO Scan Computers. “Such cutting-edge GPU-accelerated hardware will further their goals in CMAC, and Scan's team will be on-hand with hardware optimization, network configuration, and data science advice, to ensure maximum utilization of the Nvidia DGX supercomputer continues as workloads evolve."
AI-dedicated supercomputers are popping up all over the globe. This year so far has seen the South Korean Science Ministry deploy one at the Korea University's Seoul campus, Brazilian state-controlled oil and gas company Petrobras deploy another at its Cenpes center, and New York DFS announce plans for its own supercomputer in order to better understand and regulate the technology.