Loading...
Please wait, while we are loading the content...
Similar Documents
Architectural specialization for inter-iteration loop dependence patterns (2014)
| Content Provider | CiteSeerX |
|---|---|
| Author | Ilbeyi, Berkin Tan, Mingxing Liu, Gai Srinath, Shreesha Batten, Christopher Zhang, Zhiru |
| Abstract | Abstract—Hardware specialization is an increasingly com-mon technique to enable improved performance and energy ef-ficiency in spite of the diminished benefits of technology scal-ing. This paper proposes a new approach called explicit loop specialization (XLOOPS) based on the idea of elegantly en-coding inter-iteration loop dependence patterns in the instruc-tion set. XLOOPS supports a variety of inter-iteration data-and control-dependence patterns for both single and nested loops. The XLOOPS hardware/software abstraction requires only lightweight changes to a general-purpose compiler to gen-erate XLOOPS binaries and enables executing these binaries on: (1) traditional microarchitectures with minimal performance impact, (2) specialized microarchitectures to improve perfor-mance and/or energy efficiency, and (3) adaptive microarchitec-tures that can seamlessly migrate loops between traditional and specialized execution to dynamically trade-off performance vs. energy efficiency. We evaluate XLOOPS using a vertically inte-grated research methodology and show compelling performance and energy efficiency improvements compared to both simple and complex general-purpose processors. I. |
| File Format | |
| Publisher Date | 2014-01-01 |
| Access Restriction | Open |
| Content Type | Text |
| Resource Type | Article |
