The Ultimate DataFlow for Ultimate SuperComputers-on-a-Chip, for Scientific Computing, Geo Physics, Complex Mathematics, and Information Processing

Autor: Milutinovic, Veljko, Azer, Erfan Sadeqi, Yoshimoto, Kristy, Klimeck, Gerhard, Djordjevic, Miljan, Kotlar, Milos, Bojovic, Miroslav, Miladinovic, Bozidar, Korolija, Nenad, Stankovic, Stevan, Filipović, Nenad, Babovic, Zoran, Kosanic, Miroslav, Tsuda, Akira, Valero, Mateo, De Santo, Massimo, Neuhold, Erich, Skoručak, Jelena, Dipietro, Laura, Ratkovic, Ivan
Rok vydání: 2020
Předmět:
Druh dokumentu: Working Paper
Popis: This article starts from the assumption that near future 100BTransistor SuperComputers-on-a-Chip will include N big multi-core processors, 1000N small many-core processors, a TPU-like fixed-structure systolic array accelerator for the most frequently used Machine Learning algorithms needed in bandwidth-bound applications and a flexible-structure reprogrammable accelerator for less frequently used Machine Learning algorithms needed in latency-critical applications.
Databáze: arXiv