The event is over. To access the presenters’ slides, please follow the links to their talks in the schedule below.

After earlier waves of research on processing in memory (PIM) died out, the past several years have seen a resurgence of interest in PIM. Indeed, intelligent memory was one of the research topics called out in the Semiconductor Research Corporation’s JUMP 1.0 (2017) and JUMP 2.0 (2022) calls for center proposals, and is highlighted as an important area of research in the SRC-SIA Decadal Plan. This has led to various proposals for PIM architectures and several commercial products from Samsung, SK Hynix, UPMEM, and others. However, PIM still retains some stigma from the failure to influence the industry of earlier decades, and the research on PIM remains primarily concentrated in a relatively small number of universities.

To broaden the set of researchers in this topic area, we are holding a full-day, hybrid tutorial-workshop format. The tutorial part will present a background for those just beginning to be interested in the topic of PIM. The workshop part will showcase emerging research and “lessons learned” from prior PIM research, and facilitate discussions among current and future researchers in this topic area.

This event is organized by the DARPA/SRC Center for Research on Intelligent Storage and Processing in Memory (CRISP).


Final Program

8am Opening remarksKevin Skadron, University of Virginia(+)
8:10am Keynote 1: Ameen Akel, Micron Technologies
9am Guest Lectures I
Historical perspective: Why PIM now and what happened in the past
– Kevin Skadron, University of Virginia
Processing Using Memory
– Saugata Ghose, University of Illinois
Rethinking control, access, communication, and partitioning techniques for PIM-based accelerators
– Marzieh Lenjani, Apple(*)
10:30am Break
11am Panel I: Industry Perspectives
Moderator: Samira Khan, Google and University of Virginia(+)
Ameen Akel, Micron Technologies
Nuwan Jayasena, Advanced Micro Devices
Byeongho Kim, Samsung
Noon Lunch Break
1:10pm Keynote 2: Josep Torrellas, University of Illinois
2pm Guest Lectures II
Bank-level in-memory processing
– Anand Sivasubramaniam, Penn State University
Accelerating visual analytics across the memory and storage stack
– Vijay Narayanan, Penn State University
Spectrum of in-memory computing
– Reetu Das, University of Michigan
3:30pm Break
4pm Panel II: Academic Perspectives
Moderator: Ameen Akel, Micron Technologies
Samira Khan, Google and University of Virginia
Reetu Das, University of Michigan(+)
Josep Torrellas, University of Illinolis
Peter Kogge, University of Notre Dame
5pm Closing remarksKevin Skadron, University of Virginia

(*) Dr. Lenjani’s presentation is based on her Ph.D. work at the University of Virginia and does not represent her work at Apple.
(+) Replacing José Martínez, who is unable to attend.