Distributed Data Structures (pContainers)

A pContainer is the parallel equivalent of an STL container. Its data is partitioned and distributed across the machine, but the User is offered a shared object view. The pContainer distribution across the machine can be user specified or automatically selected by STAPL.

Currently, STAPL provides STL equivalent containers: pVector, pList, pArray, pMap, pSet, pMultiSet, pMultiMap, pHashMap, pMatrix and pGraph.

Views

Views are the STAPL equivalent of STL iterators in the sense that they provide a generic mechanism to access the data of the pContainers. Views emphasize processing data ranges over accessing single items. Each View may be partitioned into subviews hierarchically and this allow to adjust the degree of parallelism to the application needs and nested parallelsm. Views provide iterators to the data items to access single pContainer elements. The partition of the data and the iteration order of the elements may be independent from the physical data distribution, thus allowing easier specification of the parallel algorithms.

Parallel Algorithms (pAlgorithms)

A pAlgorithm is the parallel equivalent of an STL algorithm. A pAlgorithm is written in terms of Views operations. The hierarchical structure of input Views and the algorithm access pattern decide the degree of parallelism available for computation. A pAlgorithm can modify input views for optimized data access and/or easier algorithm specification.

Because performance of parallel algorithms is sensitive to system architecture (latency, topology, etc.) and to application data (data type, distribution, density, etc.), STAPL is designed to continually adapt to the system and the data at all levels -- from selecting the most appropriate algorithmic implementation to balancing communication granularity with latency by self-tuning the message aggregation factor, etc.(see Framework for Adaptive Algorithm Selection in STAPL)

pRange

STAPL pAlgorithms are specified as a pRange. A pRange is a Task Dependence Graph (TDG), where a Task is defined as a set of subviews representing data and a function defined over those subviews. Data dependencies are specified among Tasks to guarantee correctness of execution.

To allow an easy specification of a TDG, STAPL provides a library of dependence graphs and the capability of composing TDGs to build larger and more complex computations. The Developer can still build new TDGs directly through enumeration or closed form description of the dependencies when the dependence pattern of the computation doesn't match one of the dependence graphs provided.

Run-time System and ARMI

STAPL's run-time system (RTS) provides to the Developer and Specialist the following facilities:

Applications Developed with STAPL

STAPL is being refined and extended through the concurrent development of domain specific libraries for particle transport computations (computational physics), protein folding simulation (computational biology), and ray tracing (computational geometry). All these applications, like many problems in the physical and life sciences, make heavy use of dynamic linked data structures (e.g., graphs). STAPL provides built-in support for such irregular data structures.

Particle Transport in PDT	Protein Folding	Seismic Ray Tracing

If you have questions about STAPL, please email stapl-support@tamu.edu

ARMI: A High Level Communication Library for STAPL, Nathan Thomas, Steven Saunders, Tim Smith, Gabriel Tanase, Lawrence Rauchwerger, Parallel Processing Letters, 16(2):261-280, Jun 2006.
Journal(ps, pdf, abstract)

Parallel Protein Folding with STAPL, Shawna Thomas, Gabriel Tanase, Lucia K. Dale, Jose M. Moreira, Lawrence Rauchwerger, Nancy M. Amato, Concurrency and Computation: Practice and Experience, 17(14):1643-1656, Dec 2005.
Journal(ps, pdf, abstract)

Associative Parallel Containers In STAPL, Gabriel Tanase, Chidambareswaran (Chids) Raman, Mauro Bianco, Nancy M. Amato, Lawrence Rauchwerger, In Wkshp. on Lang. and Comp. for Par. Comp. (LCPC), Oct 2007.
Proceedings(pdf, abstract)

The STAPL pArray, Gabriel Tanase, Mauro Bianco, Nancy M. Amato, Lawrence Rauchwerger, In Proc. of Workshop MEDEA, pp. 81-88, Brasov, Romania, Sep 2007.
Proceedings(pdf, abstract)

Design and Use of htalib – a Library for Hierarchically Tiled Arrays, Ganesh Bikshandy, Jia Guo, Christoph von Praun, Gabriel Tanase, Basilio Fraguela, Maria Jesus Garzaran, David Padua, Lawrence Rauchwerger, In Wkshp. on Lang. and Comp. for Par. Comp. (LCPC), Nov 2006.
Proceedings(pdf, abstract)

A Framework for Adaptive Algorithm Selection in STAPL, Nathan Thomas, Gabriel Tanase, Olga Tkachyshyn, Jack Perdue, Nancy M. Amato, Lawrence Rauchwerger, In Proc. ACM SIGPLAN Symp. Prin. Prac. Par. Prog. (PPOPP), pp. 277-288, Chicago, Illinois, Jun 2005.
Proceedings(ps, pdf, abstract)

Parallel Protein Folding with STAPL, Shawna Thomas, Nancy M. Amato, In Proc. IEEE Int. Wkshp. on High Performance Computational Biology, Santa Fe, NM, Apr 2004.
Proceedings(ps, pdf, abstract)

ARMI: An Adaptive, Platform Independent Communication Library, Steven Saunders, Lawrence Rauchwerger, In Proc. ACM SIGPLAN Symp. Prin. Prac. Par. Prog. (PPOPP), pp. 12, San Diego, CA, Jun 2003.
Proceedings(ps, pdf, abstract)

A parallel communication infrastructure for STAPL, Steven Saunders, Lawrence Rauchwerger, In Wkshp. on Perf. Opt. for High-Level Lang. and Lib. (POHLL), New York, NY, Jun 2002.
Proceedings(ps, pdf, ppt, abstract)

STAPL: An Adaptive, Generic Parallel C++ Library, Ping An, Alin Jula, Silvius Rus, Steven Saunders, Tim Smith, Gabriel Tanase, Nathan Thomas, Nancy Amato, Lawrence Rauchwerger, In Wkshp. on Lang. and Comp. for Par. Comp. (LCPC), pp. 193-208, Cumberland Falls, Kentucky, Aug 2001.
Proceedings(ps, pdf, abstract)

STAPL: A Standard Template Adaptive Parallel C++ Library, Ping An, Alin Jula, Silvius Rus, Steven Saunders, Tim Smith, Gabriel Tanase, Nathan Thomas, Nancy Amato, Lawrence Rauchwerger, In Int. Wkshp on Adv. Compiler Technology for High Perf. and Embedded Processors, pp. 10, Bucharest, Romania, Jul 2001.
Proceedings(ps, pdf, abstract)

Standard Templates Adaptive Parallel Library (STAPL), Lawrence Rauchwerger, Francisco Arzu, K Ouchi, In Wkshp. on Lang. Comp. and Run-time Sys. for Scal. Comp. (LCR)., pp. 402-410, LNCS 1511. Pittsburgh, PA, May 1998.
Proceedings(ps, pdf, abstract)

Object Oriented Abstractions for Communication in Parallel Programs, Steven Saunders, Masters Thesis, Parasol Laboratory, Department of Computer Science, Texas A&M University, College Station, TX, May 2003.
Masters Thesis(ps, pdf, abstract)