8298

Synthesizing Structured Traversals from Attribute Grammars

Leo A. Meyerovich, Matthew E. Torok, Eric Atkinson, Rastislav Bodik
University of California, Berkeley
University of California, Berkeley, 2012
@article{meyerovich2012synthesizing,

   title={Synthesizing Structured Traversals from Attribute Grammars},

   author={Meyerovich, L.A. and Torok, M.E. and Atkinson, E. and Bod{i}k, R.},

   year={2012}

}

Download Download (PDF)   View View   Source Source   

466

views

We examine how to automatically decompose a program into structured parallel traversals over trees. In our system, programs are declaratively specified as attribute grammars and parallel traversals are defined by a compiler designed to optimize them for both GPUs and multicore CPUs. Our synthesizer automatically finds a correct schedule of the attribute grammar as structured traversals. The combination of traversals impacts algorithm performance and software integration. We therefore introduce a declarative language of traversal schedules where programmers may sketch any part of the schedule and the synthesizer will fill in the rest. For the same motivations, the synthesizer autotunes over any underconstrained fragment and can answer debugging queries about if and how such sketches can be completed. This paper presents our synthesizer, and its support for finding, specifying, debugging, and autotuning over schedules of structured tree traversals. We evaluate our approach with two case studies. First, we present a parallel schedule for a large fragment of CSS and report a speedup of 3x on multicore hardware. Second, we created a GPU-accelerated visualization language for real-time interactive animations of over 100,000 nodes.
VN:F [1.9.22_1171]
Rating: 0.0/5 (0 votes cast)

* * *

* * *

Like us on Facebook

HGPU group

141 people like HGPU on Facebook

Follow us on Twitter

HGPU group

1220 peoples are following HGPU @twitter

Featured events

* * *

Free GPU computing nodes at hgpu.org

Registered users can now run their OpenCL application at hgpu.org. We provide 1 minute of computer time per each run on two nodes with two AMD and one nVidia graphics processing units, correspondingly. There are no restrictions on the number of starts.

The platforms are

Node 1
  • GPU device 0: AMD/ATI Radeon HD 5870 2GB, 850MHz
  • GPU device 1: AMD/ATI Radeon HD 6970 2GB, 880MHz
  • CPU: AMD Phenom II X6 @ 2.8GHz 1055T
  • RAM: 12GB
  • OS: OpenSUSE 13.1
  • SDK: AMD APP SDK 2.9
Node 2
  • GPU device 0: AMD/ATI Radeon HD 7970 3GB, 1000MHz
  • GPU device 1: nVidia GeForce GTX 560 Ti 2GB, 822MHz
  • CPU: Intel Core i7-2600 @ 3.4GHz
  • RAM: 16GB
  • OS: OpenSUSE 12.2
  • SDK: nVidia CUDA Toolkit 6.0.1, AMD APP SDK 2.9

Completed OpenCL project should be uploaded via User dashboard (see instructions and example there), compilation and execution terminal output logs will be provided to the user.

The information send to hgpu.org will be treated according to our Privacy Policy

HGPU group © 2010-2014 hgpu.org

All rights belong to the respective authors

Contact us: