Coursera Flash Sale
40% Off Coursera Plus for 3 Months!
Grab it
Watch a technical lecture exploring the evolution of GPU services and accelerator-centric system design, presented by Mark Silberstein at SPCL_Bcast #53. Dive into the transformation of GPUs from gaming processors to AI powerhouses, examining their current limitations as co-processors dependent on host CPUs. Learn about innovative research conducted since 2013 by the Accelerated Computing Systems Group, focusing on alternative system designs that enable GPUs to directly access files, storage devices, SmartNICs, and network services without CPU intervention. Understand the key principles of accelerator-centric design, its benefits for programming simplification and performance optimization, and gain insights into future trends in GPU computing architecture. The hour-long presentation, recorded at ETH Zurich, offers valuable perspectives on advancing GPU capabilities beyond traditional co-processor constraints.
