What question did this study set out to answer?

The aim is to investigate how PIM can combat memory bottlenecks to enhance performance in bandwidth-intensive applications.

August 22, 2021

Aquabolt-XL: Samsung HBM2-PIM with in-memory processing for ML accelerators and beyond

Key Points

The aim is to investigate how PIM can combat memory bottlenecks to enhance performance in bandwidth-intensive applications.
Examine the limitations of traditional methods to boost memory bandwidth.
Explore PIM as a solution to enhance computing efficiency.
Analyze its impact on reducing data movement between computing and memory.
PIM significantly enhances performance for bandwidth-intensive workloads.
Improvements in energy efficiency are observed by reducing memory-related data movement.

Abstract

Using PIM to overcome memory bottleneck • Although various bandwidth increase methods have been proposed, it is physically impossible to achieve a breakthrough increase. - Limited by # of PCB wires, # of CPU ball, and thermal constraints • PIM has been proposed to improve performance of bandwidth-intensive workloads and improve energy efficiency by reducing computing-memory data movement.

Aquabolt-XL: Samsung HBM2-PIM with in-memory processing for ML accelerators and beyond

Key Points

Abstract

Cite This Study

Also Consider

Also Consider