19 May 2022

AI hardware thermal managgement: Cooling optimization for a deep learning machine

Live Webinar


Many different applications we see in our daily lives – virtual assistants like Siri and Alexa, facial recognition, self-driving cars heavily rely on deep learning and natural language processing.


Deep learning is emerging as one of the most important developments of this decade. It uses more neural network layers to simulate the human brains’ ability to learn. Using deep learning technologies, computers can be trained to accomplish specific tasks by processing large amounts of data and recognizing patterns in the data.


A chip designed specifically for deep learning offers significant advantages over a network of computers – more cores, optimized for AI, huge on-chip memory giving very low latency for lightning-fast processing. This dissipates a lot of heat that must be removed to ensure overall performance and avoid premature failures.


Join this webinar to learn how Electronic Cooling Solutions Inc – experts in thermal management- collaborated with Cerebras engineers to develop the cooling solution for the biggest chip ever built.



Guy Wagner from Electronics Cooling Solutions will address:

  • Why liquid cooling was chosen instead of air cooling for this application.
  • How the cold plate was designed using Simcenter Flotherm XT.
  • How the overall design ensures each IC operates at close to the same target temperature, which is critical for optimal performance.
  • The microchannel design that ensures each IC on the wafer gets the correct coolant flow.
  • The use of a simulation and test strategy to underpin the fidelity of the simulations.




Live webinar: AI hardware thermal management: Cooling optimization for a deep learning machine


Thursday the 19th of May 2022