Composite Cores: Pushing Heterogeneity Into a Core
MICRO-45 Proceedings of the 2012 45th Annual IEEE/ACM International Symposium on Microarchitecture
| Hi-index | 0.00 |
Over the past four decades, the number of transistors on a chip has increased exponentially in accordance with Moore's law. This has led to progress in diversified computing applications, such as health care, education, security, and communications. A number of societal projections and industrial roadmaps are driven by the expectation that these rates of improvement will continue, but the impediments to growth are more formidable today than ever before. The largest of these barriers is related to energy and power dissipation, and it is not an exaggeration to state that developing energy-efficient solutions is critical to the survival of the semiconductor industry. Extensions of today's solutions can only go so far, and without improvements in energy efficiency they are in danger of running out of steam. When examining the history of computers, a pattern emerges: successive generations of technologies, ranging from vacuum tubes to bipolar to NMOS-based technologies, were replaced when their energy overheads became prohibitive. However, there is no clear successor to todays technology, CMOS. The available alternatives are far from being commercially viable, and none has gained sufficient traction, or provided the economic justification for overthrowing the large investments made in today's CMOS-based infrastructure. Therefore, there is a strong case supporting the position that solutions to the power conundrum must come from enhanced devices, design styles and architectures, rather than a reliance on the promise of radically new technologies becoming commercially viable. This dissertation proposes that the solution to this energy crisis is the universal application of aggressive low voltage operation across all computation platforms. This can be accomplished by targeting so-called “near-threshold computing” (NTC) and by proposing novel methods to overcome the barriers that have historically relegated ultra-low voltage operation to niche markets. In particular, this dissertation explores the performance barrier that prevents the widespread adoption of NTC and provides architectures for single-core, multi-core, many-core, and digital signal processing systems. As a final proof of concept a 3D integrated prototype of 128 ARM Cortex-M3 cores and 256MB of DRAM is presented that shows a 6.4X improvement in energy-efficiency over the Intel Atom processor.