Researchers from ETH Zürich, CISPA, and New York University have published a technical paper titled “Terracotta: Enabling the Adoption of New DRAM Techniques via a Flexible DRAM Interface and Memory Controller.”
The authors state that DRAM continues to limit the performance, energy efficiency, and robustness of modern systems. Many prior works have proposed DRAM techniques that enable in-DRAM computation, improve memory access latency and parallelism, and enhance DRAM maintenance and reliability.
Problem: rigid DRAM interface
According to the paper, adopting each new DRAM technique typically requires repeated modifications to the rigid DRAM interface and memory controller. These repeated changes hinder deployment of such techniques in real systems.
The researchers observe that the DRAM commands and memory controller structures used by many different techniques are often similar. Their key idea is to exploit these similarities to compose a set of primitives that can be used to implement diverse DRAM techniques.
The Terracotta framework
To this end, they propose Terracotta, a new framework with two flexible components:
- Custom command extensions – allow DRAM vendors to define new commands within a single, standardized interface;
- Programmable memory controller – enables system designers to program support for new DRAM techniques post-silicon.
According to the paper, these two components together enable deployment of new techniques by configuring the memory controller instead of modifying the interface and controller themselves.
The work is authored by Harsh Songara, Konstantinos Kanellopoulos, F. Nisa Bostancı, Konstantinos Marios Sgouras, Ataberk Olgun, İsmail Emir Yüksel, Andreas Kosmas Kakolyris, A. Giray Yağlıkçı, and Onur Mutlu, and appears as arXiv preprint arXiv:2610.06475. DOI link.
Source: Semiconductor Engineering
