Скачать или смотреть ISCA'22 - Session 7B - Learning II and Security III

ISCA'22: The 49th International Symposium on Computer Architecture

Session 7B: Learning II and Security III
Session Chair: Martin Maas (Google)

Papers Presented:

1. A Software-defined Tensor Streaming Multiprocessor for Large-Scale Machine Learning
Authors: Dennis Abts, Garrin Kimmell, Andrew Ling (Groq Inc.); John Kim (KAIST); Matt Boyd, Andrew Bitar, Sahil Parmar, Ibrahim Ahmed, Roberto DiCecco, David Han, John Thompson, Michael Bye, Jennifer Hwang, Jeremy Fowers, Peter Lillian, Ashwin Murthy, Elyas Mehtabuddin, Chetan Tukur, Thomas Sohmers, Kris Kang, Jonathan Ross (Groq Inc.)
Link:    • ISCA'22 - Session 7B - Learning II and Sec...  

2. Themis: A Network Bandwidth-Aware Collective Scheduling Policy for Distributed Training of DL Models
Authors: Saeed Rashidi, William Won (Georgia Institute of Technology); Sudarshan Srinivasan (Intel); Srinivas Sridharan (Meta); Tushar Krishna (Georgia Institute of Technology)
Link:    • ISCA'22 - Session 7B - Learning II and Sec...  

3. MOESI-prime: Preventing Coherence-Induced Hammering in Commodity Workloads
Authors: Kevin Loughlin (University of Michigan); Stefan Saroiu, Alec Wolman (Microsoft); Yatin A. Manerkar, Baris Kasikci (University of Michigan)
Link:    • ISCA'22 - Session 7B - Learning II and Sec...  

4. PACMAN: Attacking ARM Pointer Authentication with Speculative Execution
Authors: Joseph Ravichandran, Weon Taek Na, Jay Lang, Mengjia Yan (MIT)
Link:    • ISCA'22 - Session 7B - Learning II and Sec...  

0:00 A Software-defined Tensor Streaming Multiprocessor for Large-Scale Machine Learning
26:44 Themis: A Network Bandwidth-Aware Collective Scheduling Policy for Distributed Training of DL Models
51:35 MOESI-prime: Preventing Coherence-Induced Hammering in Commodity Workloads
1:11:57 PACMAN: Attacking ARM Pointer Authentication with Speculative Execution