*banner
 

Malleable Dataflow Specification: An Essential Ingredient for Resource-Scalable Implementations
Soheil Ghiasi

Citation
Soheil Ghiasi. "Malleable Dataflow Specification: An Essential Ingredient for Resource-Scalable Implementations". Talk or presentation, 16, February, 2011; Poster presented at the Ninth Biennial Ptolemy Miniconference, Berkeley, CA.

Abstract
Variants of dataflow specification models are widely used to synthesize streaming applications for distributed-memory parallel processors. We argue that practice of specifying streaming applications using structurally-rigid dataflow models, implicitly prohibits a number of platform oriented transformations and hence, has limited portability and scalability with respect to platform resources, such as the number of its processors. We motivate Functionally-cOnsistent stRucturally-MalLEabe Streaming Specification, dubbed FORMLESS, which refers to raising the abstraction level beyond fixed-structure dataflow to address its portability and scalability limitations. To demonstrate the promise of the idea, we develop a design space exploration scheme to customize the application model for the target platform before synthesizing software. Experiments with several common streaming applications, implemented on soft multiprocessors, demonstrate improved portability and scalability over conventional dataflow specification models, and highlight practical benefits of malleable specifications.

Electronic downloads

Citation formats  
  • HTML
    Soheil Ghiasi. <a
    href="http://chess.eecs.berkeley.edu/pubs/820.html"><i>Malleable
    Dataflow Specification: An Essential Ingredient for
    Resource-Scalable Implementations</i></a>, Talk
    or presentation,  16, February, 2011; Poster presented at
    the <a
    href="http://ptolemy.eecs.berkeley.edu/conferences/11"
    >Ninth Biennial Ptolemy Miniconference</a>,
    Berkeley, CA.
  • Plain text
    Soheil Ghiasi. "Malleable Dataflow Specification: An
    Essential Ingredient for Resource-Scalable
    Implementations". Talk or presentation,  16, February,
    2011; Poster presented at the <a
    href="http://ptolemy.eecs.berkeley.edu/conferences/11"
    >Ninth Biennial Ptolemy Miniconference</a>,
    Berkeley, CA.
  • BibTeX
    @presentation{Ghiasi11_MalleableDataflowSpecificationEssentialIngredientFor,
        author = {Soheil Ghiasi},
        title = {Malleable Dataflow Specification: An Essential
                  Ingredient for Resource-Scalable Implementations},
        day = {16},
        month = {February},
        year = {2011},
        note = {Poster presented at the <a
                  href="http://ptolemy.eecs.berkeley.edu/conferences/11"
                  >Ninth Biennial Ptolemy Miniconference</a>,
                  Berkeley, CA.},
        abstract = {Variants of dataflow specification models are
                  widely used to synthesize streaming applications
                  for distributed-memory parallel processors. We
                  argue that practice of specifying streaming
                  applications using structurally-rigid dataflow
                  models, implicitly prohibits a number of platform
                  oriented transformations and hence, has limited
                  portability and scalability with respect to
                  platform resources, such as the number of its
                  processors. We motivate Functionally-cOnsistent
                  stRucturally-MalLEabe Streaming Specification,
                  dubbed FORMLESS, which refers to raising the
                  abstraction level beyond fixed-structure dataflow
                  to address its portability and scalability
                  limitations. To demonstrate the promise of the
                  idea, we develop a design space exploration scheme
                  to customize the application model for the target
                  platform before synthesizing software. Experiments
                  with several common streaming applications,
                  implemented on soft multiprocessors, demonstrate
                  improved portability and scalability over
                  conventional dataflow specification models, and
                  highlight practical benefits of malleable
                  specifications. },
        URL = {http://chess.eecs.berkeley.edu/pubs/820.html}
    }
    

Posted by Christopher Brooks on 18 Feb 2011.
Groups: ptolemy
For additional information, see the Publications FAQ or contact webmaster at chess eecs berkeley edu.

Notice: This material is presented to ensure timely dissemination of scholarly and technical work. Copyright and all rights therein are retained by authors or by other copyright holders. All persons copying this information are expected to adhere to the terms and constraints invoked by each author's copyright.

©2002-2018 Chess