Parafrase-2: An Environment for Parallelizing, Partitioning, Synchronizing, and Scheduling Programs on Multiprocessors

Polychronopoulos, Constantine D.; Girkar, Milind; Haghighat, Mohammad R.; Lee, Chia Ling; Leung, Bruce; Schouten, Dale

doi:10.1142/s0129053389000044

Cited by 162 publications

(49 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Examples include restructuring tools (such as Parafrase-2 [13], Polaris [2], and KAP [11]) and system vendors' backend compilers. This conventional approach makes the development of the frontend and the backend independent and modular.…”

Section: Motivation For the Frontend-backend Integrationmentioning

confidence: 99%

“…The HTG is a hierarchical control flow graph (HCFG) overlayed with hierarchical data-and control-dependence graphs (HDDG and HCDG). The HTG has been successfully used both in a frontend parallelizer [13] and in a backend compiler [12]. In the HTG, hierarchical nodes capture the hierarchy of program statements, and hierarchical dependence edges represent the dependence structure between tasks at the corresponding level of hierarchy.…”

Section: The Promis Irmentioning

confidence: 99%

“…Ever since the benefits of symbolic analysis were first demonstrated for compilers [13], many commercial and research compilers have adopted the use of symbolic analysis. The number of analysis and transformation techniques using symbolic analysis has increased greatly, due to the symbolic analysis capabilities of modern compilers.…”

Section: Symbolic Analysismentioning

confidence: 99%

See 2 more Smart Citations

The Design of the PROMIS Compiler

Saito

Stavrakos

Carroll

et al. 1999

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. PROMIS is a multilingual, parallelizing, and retargetable compiler with an integrated frontend and backend operating on a single unified/universal intermediate representation. This paper describes the organization and the major features of the PROMIS compiler. PROMIS exploits multiple levels of static and dynamic parallelism, ranging from task-and loop-level parallelism to instruction-level parallelism, based on a target architecture description. The frontend and the backend are integrated through a unified internal representation common to the high-level, the low-level, and the instruction-level analyses and transformations. The unified internal representation propagates hard to compute dependence information from the semantic rich frontend through the backend down to the code generator. Based on conditional algebra, the symbolic analyzer provides control sensitive and interprocedural information to the compiler. This information is used by other analysis and transformation passes to achieve highly optimized code. Symbolic analysis also helps statically quantify the effectiveness of transformations. The graphical user interface assists compiler development as well as application performance tuning.

show abstract

Section: Motivation For the Frontend-backend Integrationmentioning

confidence: 99%

Section: The Promis Irmentioning

confidence: 99%

See 1 more Smart Citation

The Design of the PROMIS Compiler

Saito

Stavrakos

Carroll

et al. 1999

Lecture Notes in Computer Science

View full text Add to dashboard Cite

show abstract

“…At the same time, NthLib communicates with the operating system to enable application adaptivity to the runtime conditions. Automatic parallelization was performed with HPF-like directives and the Parafrase-2 compiler [1,12]. Three versions of each benchmark were used in the experiments: (a) a loop parallel version produced by the native SGI parallelizing compiler and exploiting a single level of parallelism from loops; (b) a autoscheduled version, which exploits multiple levels of loop and functional parallelism; and (c) a partitioned autoscheduled version which uses autoscheduling enhanced with partitioning and clustering.…”

Section: Evaluation Frameworkmentioning

confidence: 99%

Enhancing the performance of autoscheduling in Distributed Shared Memory multiprocessors

Nikolopoulos

Polychronopoulos

Papatheodorou

1998

Euro-Par’98 Parallel Processing

View full text Add to dashboard Cite

Abstract. Autoscheduling is a parallel program compilation and execution model that combines uniquely three features: Automatic extraction of loop and functional parallelism at any level of granularity, dynamic scheduling of parallel tasks, and dynamic program adaptability on multiprogrammed shared memory multiprocessors. This paper presents a technique that enhances the performance of autoscheduling in Distributed Shared Memory (DSM) multiprocessors, targetting mainly at medium and large scale systems, where poor data locality and excessive communication impose performance bottlenecks. Our technique partitions the application Hierarchical Task Graph and maps the derived partitions to clusters of processors in the DSM architecture. Autoscheduling is then applied separately for each partition to enhance data locality and reduce communication costs. Our experimental results show that partitioning achieves remarkable performance improvements compared to a standard autoscheduling environment and a commercial parallelizing compiler.

show abstract

“…We have modi ed the Parafrase-2 source-to-source translator [12] to compile ZPL programs to SPMD C code with calls to machine-speci c communication routines. We rely on each machine's native C compiler to complete the translation to machine code.…”

Section: The Zpl Compilermentioning

confidence: 99%

SIMPLE performance results in ZPL

Lin

Snyder

1995

Languages and Compilers for Parallel Computing

View full text Add to dashboard Cite

Abstract. This paper presents performance results for ZPL programs running on the Kendall Square Research KSR-2 and the Intel Paragon. Because ZPL is a data parallel language based on the Phase Abstractions programming model, these results complement earlier claims that the Phase Abstractions model can lead to portability across MIMD computers. The ZPL language and selected aspects of the compilation strategy are brie y described, and performance results are compared against hand-coded programs.

show abstract

Parafrase-2: An Environment for Parallelizing, Partitioning, Synchronizing, and Scheduling Programs on Multiprocessors

Cited by 162 publications

References 0 publications

The Design of the PROMIS Compiler

The Design of the PROMIS Compiler

Enhancing the performance of autoscheduling in Distributed Shared Memory multiprocessors

SIMPLE performance results in ZPL

Contact Info

Product

Resources

About