SambaNova to Offer DataScale Platform to Argonne National Lab

SambaNova Systems will deliver its DataScale system to the Argonne National Laboratory to help accelerate artificial intelligence utilization.

DataScale will be used to accelerate AI for large-scale data imaging, large language models and other scientific workloads. The solution will be made available through the Argonne Leadership Computing Facility AI Testbed, a platform that allows researchers to explore deep learning and AI model workloads, SambaNova Systems said.

Researchers at Argonne have been using SambaNova’s previous-generation platform to investigate various use cases, including edge computing, neutrino physics, cancer prediction, climate modeling and engine research. Researchers used ACLF’s SambaNova system to train machine learning models, improve neutrino signal efficiency, predict tumor response, use deep learning for weather and climate data and create computational design tools for next-generation engines.

Venkatram Vishwanath, data science team leader at ALCF, said DataScale would accelerate insights into scientific data that are being produced by simulations and experiments. Marshall Choy, senior vice president of products at SambaNova, said the DataScale delivery is a multi-year deal with Argonne that is designed to allow researchers to perform new experiments and come up with novel discoveries.

Researchers can use the ALCF AI Testbed’s SambaNova DataScale system by submitting their proposals through the ALCF Director’s Discretionary program.

SambaNova, headquartered in Palo Alto, California, offers the DataScale platform and dataflow-as-a-service to private and government clients.