As chip designers face the prospect of increasingly dark silicon, there is increased interest in incorporating energy-efficient specialized coprocessors into general-purpose designs. For specialization to be a viable means of leveraging dark silicon, it must provide energy savings over the majority of execution for large, diverse workloads, and this will require deploying coprocessors in large numbers. Recent work has shown that automatically-generated application-specific coprocessors can greatly improve energy efficiency, but it is not clear that current techniques will scale to coprocessor-dominated architectures (CoDAs) with hundreds or thousands of coprocessors.We show that scaling CoDAs to include very large numbers of coprocessors is challenging because of the energy cost of interconnects, the memory system, and leakage. These overheads grow with the number of coprocessors and, left unchecked, will squander the energy gains that coprocessors can provide. The paper presents a detailed study of energy costs across a wide range of tiled CoDA designs and shows that careful choice of cache configuration, tile size, coarse-grain power management, and transistor implementation can limit the growth of these overheads. For multi-threaded workloads, designers must also take care to avoid excessive contention for coprocessors, which can significantly increase energy consumption. The results suggest that, for CoDAs that target larger workloads, amortizing shared overheads via multithreading can provide up to 3.8× reductions in energy per instruction, retaining much of the 5.3× potential of smaller designs.
Because of the constraints of space computing, the set of available processing technologies is limited. Conventionally, designers have had to choose from programmable radhard processors and fixed ASIC solutions. FPGAs provide significantly better power-performance efficiency than general purpose processors, but are more costly to program and are less flexible. For terrestrial applications, manycore processors have been adopted for a class of applications where both performance and flexible programmability are important metrics. Maestro, the first rad-hard manycore processor, has the potential to enable new capabilities for space computation. However, for many applications, certain timing-critical tasks still require the performance efficiency of an FPGA co-processor. Moreover, integrating such heterogeneous systems is challenging because the individual processing substrates have differing internal programming models. As a result, data sharing and dynamic workload scheduling across heterogeneous architectures are often suboptimal and hindered by poor scalability. The Rad-hard Unified Scalable Heterogeneous (RUSH) architecture is a heterogeneous processing platform with both a manycore chip and an FPGA. RUSH provides a unified programming model across both chips to allow for rapid development of scalable and efficient implementations. This paper overviews RUSH's technical approach and presents an example application: a WiMAX radio transceiver.
On the day my father died, I sat, for the first time, in his big leather chair behind the oversized desk in his study. I'd spent countless hours in this room with my dad, doing homework, reading books, but I had never sat behind this desk. The chair climbed up past my head as I collapsed into the soft, light brown leather. I spread my arms wide, searching for the armrests before finding them, the chair seemingly spacious enough for at least two people to comfortably sit in it. When I leaned against its back, my feet just barely touched the ground. The desk, likewise, seemed enormous from my new perspective. Made of oak, it seemed as permanent as the house itself, a shelter under which to hide from a spring thunderstorm. Matching bookshelves, the same color as the desk, lined an entire side of the room. They overflowed with a mix of three ring binders for his work as an insurance salesman and tattered, dog-eared paperbacks of the classics-Dickens, Twain, some anthologies of the Romantic poets. The shelves were so full that I thought pulling out even one of the books or binders might cause the rest to come spilling out like a dam cracking under the weight of its water. On the far wall by the door were more shelves holding a small stereo that was always tuned to NPR or the local opera station, depending on his mood that night. Pictures of my family stared back at me from distant locales-a mountain in Colorado, a golf course in Arizona, a beach in Hawai'i. The deep red color on the walls made the room feel closed-off from the world, a sanctuary for a father and his son. I looked to the far left corner of the desk, the spot I occupied as a kid when my dad and I worked together in the office. We didn't speak much during those evenings. He would clack away at his typewriter, occasionally grumbling to himself or reading aloud *** We're not invited, but walk into the party like we own the place, like rock stars, full of swagger and certainty, and stand just inside the front door. The crowd inside, a collection of khakis, North Face fleeces, and backward caps, stops its collective conversation in mid-sentence and stares us down. For a moment no one in the room so much as utters a sound, and the Coldplay on in the background serves as soundtrack to this standoff. This was a bad idea. We may be rock stars, but we've run up against 20 or 30 yuppies who probably spend a fair amount of time swimming laps and riding elliptical machines at the country club. I look around the room, avoiding the hard, piercing stares, take in the maple-wood flooring, the eggshell white walls, follow the line of the large
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.
hi@scite.ai
10624 S. Eastern Ave., Ste. A-614
Henderson, NV 89052, USA
Copyright © 2024 scite LLC. All rights reserved.
Made with 💙 for researchers
Part of the Research Solutions Family.