Compiling global name-space parallel loops for distributed execution

Koelbel, Charles; Mehrotra, Piyush

doi:10.1109/71.97901

Cited by 190 publications

(70 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…These primitives have been integrated into the Vienna Fortran Compilation System and are also being implemented in the context of the Fortran D Programming environment being developed at Rice University. Similar strategies to preprocess DO loops at runtime to extract the communication pattem have also been developed within the context of the Kali language [ 6,36 J. Explicit runtime generation of messages is also performed by other researchers [29,30,32]; however, these do not save the extracted communication pattem to avoid recalculation.…”

Section: Related Workmentioning

confidence: 99%

“…The data distribution of each of the arrays is determined at runtime; thus the compiler cannot detect which references are local and which are not. In such situations, runtime techniques such as those developed in other projects [ 6,14] are needed to generate and exploit the communication pattern.…”

Section: Irregular Distributionsmentioning

confidence: 99%

“…The Kali compiler [6] was the first system to support both regular and irregular computations, using an inspector/ executor strategy to handle indirectly distributed data. It produces code which is independent of the number of processors.…”

Section: Related Workmentioning

confidence: 99%

“…Research in compiler technology has so far resulted in the development of a number of prototype systems, such as Kali [6], SUPERB [7,8], and the MIMDizer [9]. In contrast to the current programming paradigm, these systems enable users to write code using global data references, as on a shared memory machine, but require them to specify the distribution of the program's data.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Programming in Vienna Fortran

Chapman

Mehrotra

Zima

1992

Scientific Programming

194

View full text Add to dashboard Cite

Exploiting the full performance potential of distributed memory machines requires a careful distribution of data across the processors. Vienna Fortran is a language extension of Fortran which provides the user with a wide range of facilities for such mapping of data structures. In contrast to current programming practice, programs in Vienna Fortran are written using global data references. Thus, the user has the advantages of a shared memory programming paradigm while explicitly controlling the data distribution. In this paper, we present the language features of Vienna Fortran for FoRTRAN 77, together with examples illustrating the use of these features

show abstract

Section: Related Workmentioning

confidence: 99%

Section: Irregular Distributionsmentioning

confidence: 99%

Section: Related Workmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Programming in Vienna Fortran

Chapman

Mehrotra

Zima

1992

Scientific Programming

194

View full text Add to dashboard Cite

show abstract

“…In the context of parallelizing irregular applications, gather/scatter parallelism is commonly used in irregular applications where the programmer has specified the data decomposition for a distributed array [52,33,27,68,11]. Typically there is language and compilation support for data distribution specifications, parallel loops, and reductions, which involves generating code with calls to the appropriate inspector, scheduling, and gather/scatter functions in a run-time library such as CHAOS [45].…”

Section: Data and Iteration Embedding Reorderingsmentioning

confidence: 99%

An approach for code generation in the Sparse Polyhedral Framework

et al. 2016

View full text Add to dashboard Cite

Applications that manipulate sparse data structures contain memory reference patterns that are unknown at compile time due to indirect accesses such as A [B[i]]. To exploit parallelism and improve locality in such applications, prior work has developed a number of run-time reordering transformations (RTRTs). This paper presents the Sparse Polyhedral Framework (SPF) for specifying RTRTs and compositions thereof and algorithms for automatically generating efficient inspector and executor code to implement such transformations. Experimental results indicate that the performance of automatically generated inspectors and executors competes with the performance of hand-written ones in some cases.

show abstract