SARA: Scaling a Reconfigurable Dataflow Accelerator
Authors:
- 1 June 2021
- conference paper
- Published by Institute of Electrical and Electronics Engineers (IEEE)
- p. 1041-1054
Abstract
The need for speed in modern data-intensive work-loads and the rise of "dark silicon" in the semiconductor industry are pushing for larger, faster, and more energy and area-efficient architectures, such as Reconfigurable Dataflow Accelerators (RDAs). Nevertheless, challenges remain in developing mechanisms to effectively utilize the compute power of these large-scale RDAs. To address these challenges, we present SARA, a compiler that employs a novel mapping strategy to efficiently utilize large-scale RDAs. Starting from a single-threaded imperative abstraction, SARA spatially maps a program onto RDA's distributed resources, exploiting dataflow parallelism within and across hyperblocks to saturate the compute throughput of an RDA. SARA introduces (a) compiler-managed memory consistency (CMMC), a control paradigm that hierarchically pipelines a nested and data-dependent control-flow graph onto a dataflow architecture, and (b) a compilation flow that decomposes the program graph across distributed heterogeneous resources to hide low-level RDA constraints from programmers. Our evaluation shows that SARA achieves close to perfect performance scaling on a recently proposed RDA—Plasticine. Over a mix of deep-learning, graph-processing, and streaming applications, SARA achieves a 1.9× geo-mean speedup over a Tesla V100 GPU using only 12% of the silicon area.Keywords
Funding Information
- Advanced Research Projects Agency
- National Science Foundation
This publication has 43 references indexed in Scilit:
- High-level Design Tools for Complex DSP Applications2012
- ADRES: An Architecture with Tightly Coupled VLIW Processor and Coarse-Grained Reconfigurable Matrix2003
- The Raw microprocessor: a computational fabric for software circuits and general-purpose programsIEEE Micro, 2002
- PipeRench: a reconfigurable architecture and compilerComputer, 2000
- Long Short-Term MemoryNeural Computation, 1997
- Scheduling data-flow graphs via retiming and unfoldingIEEE Transactions on Parallel and Distributed Systems, 1997
- Effective compiler support for predicated execution using the hyperblockACM SIGMICRO Newsletter, 1992
- Some scheduling techniques and an easily schedulable horizontal architecture for high performance scientific computingACM SIGMICRO Newsletter, 1981
- Design of ion-implanted MOSFET's with very small physical dimensionsIEEE Journal of Solid-State Circuits, 1974
- The Transitive Reduction of a Directed GraphSIAM Journal on Computing, 1972