Difference between revisions of "Rodinia:Accelerating Compute-Intensive Applications with Accelerators"

From Rodinia
Jump to: navigation, search
(The Rodinia Benchmark Suite)
Line 1: Line 1:
 
A vision of heterogeneous computer systems that incorporate diverse <br/>
 
A vision of heterogeneous computer systems that incorporate diverse <br/>
 
accelerators and automatically select the best computational unit for <br/>
 
accelerators and automatically select the best computational unit for <br/>
a particular task is widely shared amongresearchers and many industry <br/>
+
a particular task is widely shared among researchers and many industry <br/>
 
analysts; however, there are no agreed-upon benchmarks to support the <br/>
 
analysts; however, there are no agreed-upon benchmarks to support the <br/>
 
research needed in the development of such a platform. There are many <br/>
 
research needed in the development of such a platform. There are many <br/>
 
suites for parallel computing on general-purpose CPU architectures,<br/>
 
suites for parallel computing on general-purpose CPU architectures,<br/>
 
but accelerators fall into a gap that is not covered by current benchmark <br/>
 
but accelerators fall into a gap that is not covered by current benchmark <br/>
suites or benchmark development.  
+
development. Rodinia is released to address this concern.
  
 
===The Rodinia Benchmark Suite===
 
===The Rodinia Benchmark Suite===
  
Rodinia is designed for heterogeneous computing infrastructures, and, <br/>
+
Rodinia is designed for heterogeneous computing infrastructures with <br/>
using OpenMP and CUDA, targets both GPUs and multicore CPUs. <br/>  
+
OpenMP, OpenCL and CUDA implementations. <br/>  
They have been parallelized with OpenMP for multicore CPUs and with <br/>
+
CUDA for GPUs. We have been preparing for the OpenCL release.
+
  
 
Current Rodinia applications:
 
Current Rodinia applications:
Line 42: Line 40:
 
<TABLE border="1">
 
<TABLE border="1">
 
<TR><TH WIDTH="150">Applications<TH WIDTH="150">Dwarves<TH WIDTH="150">Domains<TH WIDTH="120">Parallel Model<TH WIDTH="400">Comment
 
<TR><TH WIDTH="150">Applications<TH WIDTH="150">Dwarves<TH WIDTH="150">Domains<TH WIDTH="120">Parallel Model<TH WIDTH="400">Comment
<TR><TD>[[SQL Database | SQLite Select]]<TD>Map Reduce<TD>Relational Database<TD ALIGN="center">CUDA<TD>This benchmark needs an OCL version and also may be too simple)
+
<TR><TD>[[SQL Database | SQLite Select]]<TD>Map Reduce<TD>Relational Database<TD ALIGN="center">CUDA<TD>This benchmark needs an OCL version and also may be too simple
 
<TR><TD>[[Cell | 3D Stencil]]<TD>Structured Grid<TD>Cellular Automation<TD ALIGN="center">CUDA<TD>Will be superseded by a more sophisticated 3D benchmark
 
<TR><TD>[[Cell | 3D Stencil]]<TD>Structured Grid<TD>Cellular Automation<TD ALIGN="center">CUDA<TD>Will be superseded by a more sophisticated 3D benchmark
 
<TR><TD>[[Hybrid Sort]]<TD>Sorting<TD>Sorting Algorithms<TD ALIGN="center">CUDA<TD>This benchmark is under consideration to include in the main suite
 
<TR><TD>[[Hybrid Sort]]<TD>Sorting<TD>Sorting Algorithms<TD ALIGN="center">CUDA<TD>This benchmark is under consideration to include in the main suite

Revision as of 17:44, 10 November 2011

A vision of heterogeneous computer systems that incorporate diverse
accelerators and automatically select the best computational unit for
a particular task is widely shared among researchers and many industry
analysts; however, there are no agreed-upon benchmarks to support the
research needed in the development of such a platform. There are many
suites for parallel computing on general-purpose CPU architectures,
but accelerators fall into a gap that is not covered by current benchmark
development. Rodinia is released to address this concern.

The Rodinia Benchmark Suite

Rodinia is designed for heterogeneous computing infrastructures with
OpenMP, OpenCL and CUDA implementations.

Current Rodinia applications:

ApplicationsDwarvesDomainsParallel ModelIncre. Ver.
LeukocyteStructured GridMedical ImagingCUDA, OMP
Heart WallStructured GridMedical ImagingCUDA, OMP
MUMmerGPUGraph TraversalBioinformaticsCUDA, OMP
CFD Solver1Unstructured GridFluid DynamicsCUDA, OMP, OCL
LU Decomposition Dense Linear AlgebraLinear AlgebraCUDA, OMP
HotSpotStructured GridPhysics SimulationCUDA, OMP, OCL
Back PropagationUnstructured GridPattern RecognitionCUDA, OMP, OCL
Needleman-WunschDynamic ProgrammingBioinformaticsCUDA, OMP, OCL
KmeansDense Linear AlgebraData MiningCUDA, OMP, OCL
Breadth-First Search1Graph TraversalGraph AlgorithmsCUDA, OMP, OCL
SRADStructured GridImage ProcessingCUDA, OMP, OCL
Streamcluster1Dense Linear AlgebraData MiningCUDA, OMP, OCL
Particle FilterStructured GridMedical ImagingCUDA, OMP
PathFinderDynamic ProgrammingGrid TraversalCUDA, OCL
Gaussian EliminationDense Linear AlgebraLinear AlgebraCUDA, OCL
k-Nearest NeighborsDense Linear AlgebraData MiningCUDA, OMP, OCL
LavaMD2Structured GridMolecular DynamicsCUDA, OMP, OCL

Other applications under evaluation:

ApplicationsDwarvesDomainsParallel ModelComment
SQLite SelectMap ReduceRelational DatabaseCUDAThis benchmark needs an OCL version and also may be too simple
3D StencilStructured GridCellular AutomationCUDAWill be superseded by a more sophisticated 3D benchmark
Hybrid SortSortingSorting AlgorithmsCUDAThis benchmark is under consideration to include in the main suite
MyocyteStructured GridBiological SimulationCUDA, OMPThis benchmark needs an OpenCL version

1Ana Lucia Varbanescu and Jianbin Fang, Delft University of Technology contributed the OpenCL version.

2Collaboration with Lawrence Livermore National Laboratory.

Application Download

Please read the license file.

Also, if your use of Rodinia results in a publication, please cite:

[1] S. Che, M. Boyer, J. Meng, D. Tarjan, J. W. Sheaffer, S.-H. Lee, and K. Skadron.
Rodinia: A Benchmark Suite for Heterogeneous Computing. In Proceedings of the IEEE
International Symposium on Workload Characterization (IISWC), pp. 44-54, Oct. 2009.

[2] S. Che, J. W. Sheaffer, M. Boyer, L. G. Szafaryn, L. Wang, and K. Skadron.
A Characterization of the Rodinia Benchmark Suite with Comparison to Contemporary
CMP Workloads. In Proceedings of the IEEE International Symposium on Workload
Characterization, Dec. 2010.

This work is supported by NSF grant nos. IIS-0612049, CNS-0916908 and CNS-0615277, a grant from the SRC under task no. 1607, and grants from AMD, NEC labs, and NVIDIA Research.