Abstract | ||
---|---|---|
To meet the conflicting goals of high-performance low-cost embedded systems, critical application loop nests are commonly executed on specialized hardware accelerators. These loop accelerators are traditionally designed in a single-function manner, wherein each loop nest is implemented as a dedicated hardware block. This paper focuses on hardware sharing across loop nests by creating multifunction loop accelerators, or accelerators capable of executing multiple algorithms. A compiler-based system for automatically synthesizing multifunction loop accelerator architectures from C code is presented. We compare the effectiveness of three architecture synthesis approaches with varying levels of complexity: sum of individual accelerators, union of individual accelerators, and joint accelerator synthesis. Experiments show that multifunction accelerators achieve substantial hardware savings over combinations of single-function designs. In addition, the union approach to multifunction synthesis is shown to be effective at creating low-cost hardware by exploiting hardware sharing, while remaining computationally tractable. |
Year | DOI | Venue |
---|---|---|
2006 | 10.1145/1176254.1176322 | CODES+ISSS |
Keywords | Field | DocType |
multifunction loop accelerator,specialized hardware accelerator,loop nest,low-cost hardware,hardware efficiency,loop accelerator,critical application loop nest,individual accelerator,multifunction loop accelerator architecture,dedicated hardware block,hardware sharing,embedded systems,high level synthesis,hardware accelerator,multifunction | Architecture,Computer science,Parallel computing,High-level synthesis,Real-time computing,Computer hardware,Embedded system | Conference |
ISBN | Citations | PageRank |
1-59593-370-0 | 14 | 0.69 |
References | Authors | |
21 | 4 |
Name | Order | Citations | PageRank |
---|---|---|---|
Kevin Fan | 1 | 335 | 20.29 |
Manjunath Kudlur | 2 | 1997 | 71.21 |
Hyunchul Park | 3 | 341 | 17.56 |
Scott Mahlke | 4 | 4811 | 312.08 |