Compute Substrate for Software 2.0

Autor: Ljubisa Bajic, Syed Zohaib Gilani, James Connolly, Tony Zhou, Milos Trajkovic, Hassan Farooq, Lejla Bajic, Daniel Rosen, Ivan Hamer, Raymond Kim, Joseph W.K. Chu, Aleksandar Cejkov, Alexander Drouillard, Allan Rui, David C. Thompson, Rakesh Shaji Lal, Joy Yu Ting Chen, Davor Capalija, Kei-Ming Kwong, Boris Drazic, Zahi Moudallal, Stanislav Sokorac, Renjith Retnamma, Ivan Matosevic, Shripad Karodi, Utku Aydonat, Emilio Munoz, Djordje Maksimovic, Namal Rajatheva, Almeet Bhullar, Keivan Dabiri, Stephen Alexander Chin, Aleksandar Knezevic, James Sun, Miles Dooley, Jasmina Vasiljevic, Armond Paiva, Kyle Mabee, Dragoljub Ignjatovic, Sean Nijjar, Charles Lee, Andrew Lewycky, Matthew Walker, Akhmed Rakhmati
Rok vydání: 2021
Předmět:
Zdroj: IEEE Micro. 41:50-55
ISSN: 1937-4143
0272-1732
Popis: The rapidly growing compute demands of AI necessitate the creation of new computing architectures and approaches. Tenstorrent designed its architecture (embodied in Grayskull and Wormhole devices) to tackle this challenge via two fundamental and synergistic approaches. The first is via compute-on-packets fabric that is built from ground up for massive scaleout. The second is the ability to execute dynamic computation, built into the compiler, runtime software and hardware architecture. By combining these approaches, TensTorrent will enable continued scaling of AI workloads.
Databáze: OpenAIRE