SHF: Large:Collaborative Research: Architecting the Next Generation Memory Hierarchy - A Holistic Approach

The memory system continues to be a major performance and power

bottleneck in nearly all computing systems. And, it is becoming

increasingly more so with major application, architecture, and

technology trends. Embedded applications that acquire and

process real-time data, Internet and cloud applications that have to

analyze large databases, and the exa-scale era HPC applications that

need to crunch voluminous data sets are just a few examples of

increasingly data-intensive applications that require high memory

capacity, performance, and energy efficiency. Thus, the well-known

memory wall problem has become even more difficult to surmount and

needs a fundamental rethinking of the memory hierarchy design for future

computing platforms.

The goal of this proposal is to fundamentally and holistically

rethink the design of the entire memory hierarchy taking into consideration

the emerging device/memory technologies and to exploit the design trade-offs

at different layers of the system stack -- from devices to micro-architecture,

compilers and runtime systems. The solution will cover innovations in

architecting and optimizing the entire memory path consisting of the caches,

on-chip networks, memory controller and main memory. The objective

is to enable 100X improvement in memory capacity over the next

decade, while providing 5X improvement in performance and

10X improvement in energy efficiency. The proposed research has the potential to transform the design of

next-generation memory systems for the multi-core era, which is expected

to be a ubiquitous part of the entire IT sector.

The cross-cutting nature of this research can foster new research directions

in several areas, spanning technology/energy-aware design, computer architecture,

compilers, and system/application software. With the memory system forming

the backbone of nearly every envisioned future application domain, the

broader impact of this research can accelerate the design and deployment

of future applications. This project will enable transfer of research

results to industry, enhance undergraduate and graduate student training

including under-represented students, and contribute to the development of new

research and teaching tools.