Paradigm Shift in Big Data SuperComputing: DataFlow vs. ControlFlow

The paper discusses the shift in the computing paradigm and the programming model for Big Data problems and applications. We compare DataFlow and ControlFlow programming models through their quantity and quality aspects. Big Data problems and applications that are suitable for implementation on DataFlow computers should not be measured using the same measures as ControlFlow computers. We propose a new methodology for benchmarking, which takes into account not only the execution time, but also the power and space, needed to complete the task. Recent research shows that if the TOP500 ranking was based on the new performance measures, DataFlow machines would outperform ControlFlow machines. To support the above claims, we present eight recent implementations of various algorithms using the DataFlow paradigm, which show considerable speed-ups, power reductions and space savings over their implementation using the ControlFlow paradigm.

[1]  Jack B. Dennis,et al.  A preliminary architecture for a basic data-flow processor , 1974, ISCA '75.

[2]  Michael J. Flynn.  Area - Time - Power and Design effort: the basic tradeoffs in Application Specific Systems , 2005, ASAP.

[3]  D. Oriato,et al.  Acceleration of a Meteorological Limited Area Model with Dataflow Engines , 2012, 2012 Symposium on Application Accelerators in High Performance Computing.

[4]  Wayne Luk,et al.  A fully-pipelined expectation-maximization engine for Gaussian Mixture Models , 2012, 2012 International Conference on Field-Programmable Technology.

[5]  Michael W. Towsey,et al.  Managing and Analysing Big Audio Data for Environmental Monitoring , 2013, 2013 IEEE 16th International Conference on Computational Science and Engineering.

[6]  Oskar Mencer.  Multiscale Dataflow Computing , 2013 .

[7]  S. Bliesner,et al.  Predictive analytics with aviation big data , 2013, 2013 Integrated Communications, Navigation and Surveillance Conference (ICNS).

[8]  Tilak Agerwala,et al.  Data Flow Systems: Guest Editors' Introduction , 1982, Computer.

[9]  Matthias Ort,et al.  BIG DATA infrastructures for pharmaceutical research , 2013, 2013 IEEE International Conference on Big Data.

[10]  Wayne Luk,et al.  A mixed precision Monte Carlo methodology for reconfigurable accelerator systems , 2012, FPGA '12.

[11]  Mateo Valero,et al.  Moving from petaflops to petadata , 2013, CACM.

[12]  Michael J. Flynn,et al.  Beyond Traditional Microprocessors for Geoscience High-Performance Computing Applications , 2011, IEEE Micro.

[13]  Wayne Luk,et al.  Hardware Acceleration of Genetic Sequence Alignment , 2013, ARC.