What is the naive aggregation of USEb(L)?

Because one iteration's use may only be exposed to the de nitions in some previous iterations of the same loop, a naive aggregation of USEb(L) may exaggerate the exposed use set.

What is the way to handle nonlinear subscripts?

Compared with the dependence analysis-based approach [12], which has to employ parametric integer programming in its most general case, their approach is more e - cient and can handle nonlinear subscripts that cannot be handled by integer programming.

What is the aggregation of do loops?

Because the authors only consider do loops, the aggregation is a relatively straightforward interpretation of loop index and boundaries in the do-entry of the loop.

Why is loop S1 not executed in parallel?

Consider the loop:Because every iteration of loop S1 accesses the same elements of array A, loop S1 cannot be executed in parallel.

What is the algorithm to use to linearize an array in the subroutine?

Their interprocedural mapping program will linearize an array in the subroutine if it has a di erent number of dimensions as in the caller.

What is the condition for privatization of array A?

For the purpose of eliminatingmemory-related dependence in this paper, the array A in the previous example need not be privatized.

What is the way to solve the ambiguities?

Some of the ambiguities can be resolved by a more powerful forward substitution algorithm than that available in their current system.

How do the authors compute the subroutine summary set?

The authors run the algorithm in bottom-up order on the program call tree, such that each time the authors encounter a subroutine call in a program, the summary set for the subroutine has already been computed.

Open AccessBook Chapter10.1007/3-540-45403-9_8

Automatic Array Privatization

Peng Tu, +1 more

- 12 Aug 1993

- pp 500-521

161

TL;DR: In this article, a technique for automatic array privatization is presented, which uses data flow analysis of array references to identify privatizable arrays intraprocedurally as well as interprocedural.

Abstract: Array privatization is one of the most effective transformations for the exploitation of parallelism. In this paper, we present a technique for automatic array privatization. Our algorithm uses data flow analysis of array references to identify privatizable arrays intraprocedurally as well as interprocedurally. It employs static and dynamic resolution to determine the last value of a lived private array. We compare the result of automatic array privatization with that of manual array privatization and identify directions for future improvement. To enhance the effectiveness of our algorithm, we develop a goal directly technique to analysis symbolic variables in the present of conditional statements, loops and index arrays.

AI Agents for this Paper

Find similar papers on Google Scholar, PubMed and Arxiv
Write a critical review of this paper
Analyze citations of this paper to find unaddressed research gaps

Most frequently asked questions

1. What are the two main types of dependence?

Anti-dependence and output dependence are memory-related or false dependence because they are not caused by the ow of values from one statement to another, but by the reuse of memory locations.

2. What is the purpose of data ow analysis?

Data ow analysis examines the ow of values through a program and solves data ow problems by propagating information along the paths of a control ow graph.

3. What is the reason why PRIb is used in the loop?

PRIb contains all the array uses that are covered by some de nition in the same iteration of the loop; some of the uses are conditional, where they are covered by some conditional de nition.

4. Why is it easy to track the value of a variable by its name?

Because in a SSA representation, each use of a variable can be reached by exactly one assignment to that variable, it is very easy to track the value of a variable by its name.

Figure 1. Algorithm for Identifying Privatizable Arrays

Citations

Journal Article•10.1145/197405.197406

Compiler transformations for high-performance computing

David F. Bacon, +2 more

- 01 Dec 1994

- ACM Computing Surveys

TL;DR: This survey is a comprehensive overview of the important high-level program restructuring techniques for imperative languages, such as C and Fortran, and describes the purpose of each transformation, how to determine if it is legal, and an example of its application.

...read moreread less

•Posted Content

Tiramisu: A Polyhedral Compiler for Expressing Fast and Portable Code

Riyadh Baghdadi, +8 more

- 27 Apr 2018

- arXiv: Programming Languages

TL;DR: Tiramisu introduces a scheduling language with novel commands to explicitly manage the complexities that arise when targeting these systems, designed for the areas of image processing, stencils, linear algebra and deep learning.

...read moreread less

245

Journal Article•10.1109/MC.2009.385

Cetus: A Source-to-Source Compiler Infrastructure for Multicores

Chirag Dave, +5 more

- 01 Dec 2009

- IEEE Computer

TL;DR: The Cetus tool provides an infrastructure for research on multicore compiler optimizations that emphasizes automatic parallelization and provides the most important parallelization passes as well as the underlying enabling techniques.

...read moreread less

229

•Proceedings Article•10.1145/224538.224648

Gated SSA-based demand-driven symbolic analysis for parallelizing compilers

Peng Tu, +1 more

- 03 Jul 1995

TL;DR: A GSA-based technique that performs more efficient and more precise symbolic analysis of predicated assignments, recurrences and index arrays by utilizing the gating predicate information embedded in the GSA and the control dependence information in the program flow graph is presented.

...read moreread less

144

•Book

Automatic array privatization

Peng Tu, +1 more

- 01 Jun 2001

TL;DR: This paper uses data flow analysis of array references to identify privatizable arrays intraprocedural as well as interprocedurally and develops a goal directly technique to analysis symbolic variables in the present of conditional statements, loops and index arrays.

...read moreread less

132

...

Expand

References

•Book

Compilers: Principles, Techniques, and Tools

Alfred V. Aho, +2 more

- 01 Jan 1986

TL;DR: This book discusses the design of a Code Generator, the role of the Lexical Analyzer, and other topics related to code generation and optimization.

...read moreread less

9.7K

•Journal Article•10.1145/24039.24041

The program dependence graph and its use in optimization

Jeanne Ferrante, +2 more

- 01 Jul 1987

- ACM Transactions on Programming Language...

TL;DR: An intermediate program representation, called the program dependence graph (PDG), that makes explicit both the data and control dependences for each operation in a program, allowing transformations to be triggered by one another and applied only to affected dependences.

...read moreread less

2.8K

•Journal Article•10.1145/115372.115320

Efficiently computing static single assignment form and the control dependence graph

Ron Cytron, +4 more

- 01 Oct 1991

- ACM Transactions on Programming Language...

TL;DR: In this article, the authors present new algorithms that efficiently compute static single assignment forms and control dependence graphs for arbitrary control flow graphs using the concept of {\em dominance frontiers} and give analytical and experimental evidence that these data structures are usually linear in the size of the original program.

...read moreread less

2.4K

•Proceedings Article•10.1145/512760.512770

Automatic discovery of linear restraints among variables of a program

Patrick Cousot, +1 more

- 01 Jan 1978

1.9K

Journal Article•10.1145/7902.7904

Advanced compiler optimizations for supercomputers

David Padua, +1 more

- 01 Dec 1986

- Communications of The ACM

TL;DR: Compilers for vector or multiprocessor computers must have certain optimization features to successfully generate parallel code to be able to operate on parallel systems.

...read moreread less

782

...

Expand

Automatic Array Privatization

Chat with Paper

AI Agents for this Paper

Most frequently asked questions

1. What are the two main types of dependence?

2. What is the purpose of data ow analysis?

3. What is the reason why PRIb is used in the loop?

4. Why is it easy to track the value of a variable by its name?

5. What is the naive aggregation of USEb(L)?

6. What is the way to handle nonlinear subscripts?

7. What is the aggregation of do loops?

8. Why is loop S1 not executed in parallel?

9. What is the algorithm to use to linearize an array in the subroutine?

10. What is the condition for privatization of array A?

11. What is the way to solve the ambiguities?

12. How do the authors compute the subroutine summary set?

Figures

Citations

Compiler transformations for high-performance computing

Tiramisu: A Polyhedral Compiler for Expressing Fast and Portable Code

Cetus: A Source-to-Source Compiler Infrastructure for Multicores

Gated SSA-based demand-driven symbolic analysis for parallelizing compilers

Automatic array privatization

References

Compilers: Principles, Techniques, and Tools

The program dependence graph and its use in optimization

Efficiently computing static single assignment form and the control dependence graph

Automatic discovery of linear restraints among variables of a program

Advanced compiler optimizations for supercomputers

Related Papers (5)

Array privatization for parallel execution of loops

Advanced compiler optimizations for supercomputers

Array expansion

Performance analysis pf parallelizing compilers on the Perfect Benchmarks programs

Global optimizations for parallelism and locality on scalable parallel machines