Data flow graph optimization in if1

Skedzielewski, S.; Welcome, M. L.

doi:10.1007/3-540-15975-4_27

Cited by 29 publications

(7 citation statements)

References 0 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The use of standard optimization s on dataflow graphs has been discussed in the literature [19] [18]. This seems like a useful area for further research, as the future of functional languages depends heavily on improved compilation strategies.…”

Section: Optimizat Ionmentioning

confidence: 99%

“…The second goal of the project was to investigate optimizatio n techniques for dataflow programs. A paper showing the optimizatio n of an intermedia te graph description language called IFl [19] demonstrat ed how a number of traditional imperative language optimizatio ns can be applied to dataflow graphs. There are some techniques that are no longer useful in the absence of side-effects , but any optimizatio n that reduces the amount of computatio n in a program is useful in both models.…”

Section: Conclus Ionsmentioning

confidence: 99%

“…The second version of the compiler, while adequate, was based on a straight-forwa rd translation with a minimum of analysis. We were intrigued by the notion of applying traditional optimization techniques to dataflow graphs, an idea that has been explored by other researchers [19]. Rather than extensively modifying the existing compiler.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Pythia: A Parallel Compiler for Delirium

Sharp¹

1990

View full text Add to dashboard Cite

Pythia is an optimizing compiler for the coordination language Delirium, written in Delirium. It is part of the Madness project, which investigates both an alternative to the traditional dataflow model and an alternative to the traditional dataflow implementation strategy. Delirium embeds imperatively defined operators within a functional context, giving the programmer control over the granularity of a computation. The application's control structure is expressed using powerful functional-langu age constructs like closures and function valued parameters. The bulk of actual computation, however, is expressed in a convenient imperative language. Pythia performs traditional optimization strategies like macro-expansio n, common-subexp ression elimination, and constant propagation.

show abstract

Section: Optimizat Ionmentioning

confidence: 99%

Section: Conclus Ionsmentioning

confidence: 99%

See 1 more Smart Citation

Pythia: A Parallel Compiler for Delirium

Sharp¹

1990

View full text Add to dashboard Cite

show abstract

“…A front-end compiler converts SISAL code into a graphical, machine-independent intermediate format, known as IF1 [21]. Several standard optimisations, such as eliminating common subexpressions and moving constants out of loops, are performed by IFl-to-IF1 transformations [22]. Machine-specific code generators convert the IF1 code into appropriate machine-code for a variety of parallel machines, including the Cray-1, the Sequent Balance, and the MDFM.…”

Section: Language Levelmentioning

confidence: 99%

Performance issues in dataflow machines

Gurd

Bohm

Teo

1987

Future Generation Computer Systems

View full text Add to dashboard Cite

Design Issues for Parallel Computers Architectural LevelsThis paper explores issues affecting the performance of a particular style of parallel computer system. In considering such issues, it is important to distinguish between different levels of the system architecture.Broadly speaking, computer systems are designed at three levels. The first, usually thought of as the lowest level, is that of the hardware architec- North-Holland Future Generations Computer Systems 3 (1987) [285][286][287][288][289][290][291][292][293][294][295][296][297] ture and its associated machine-code. This is often referred to as the machine level. The next, intermediate level is that of the so-called high-level programming language, from which machine-code is generated. This is often referred to as the language level. Most often, this is not at a sufficiently high level to express specific applications in a convenient way. Consequently, the highest level, known as the applications level, is constructed for particular application areas on top of the language level. The interface to this level is a kind of 'very-high-level', applications-oriented language.As far as the end-user of a computer system is concerned, performance at the applications level is most important. However, as this varies significantly according to the particular application, and in a somewhat unpredictable fashion, most attempts at evaluating performance are made at the two lower levels. This paper broadly conforms to this pattern, but an attempt is made to relate the performance assessment to a specific set of applications, the so-called scientific problems, which involve a substantial amount of numerical calculation. Parallel WorkloadParallel system performance is predominantly determined by design decisions made at the machine and language levels. The most critical of these decisions concern the nature, identification, distribution and control of the parallel workload.Different basic models of parallel computation can create workloads of a fundamentally different nature. For example, although many parallel programs are written as systems of communicating parallel processes, there is a critical distinction between systems that implement the communication using shared memory and those that use direct message-passing. The workload in the former involves accesses to a hardware third party (the shared memory) that are not necessary in the latter. Also, there is an important difference between such process-based schemes, which split the overall workload into coarse-grain chunks, and systems, such as dataflow, that segregate work into very fine-grain components. The former re- 0376-5075788753.50 © 1988, Elsevier Science Publishers B.V. (North-Holland)

show abstract

“…Our discussions are based on the dataflow representation of a program which is similar to IFl [8]. A program can be described as a set of dataflow graphs with each representing a function and one of them being the main function graph where the computation starts.…”

Section: The Intermediate Dataflow Graphmentioning

confidence: 99%

Efficient storage reuse of aggregates in single assignment languages

Kirkham

1996

Lecture Notes in Computer Science

View full text Add to dashboard Cite

Abstract. The storage reuse of aggregates is a key problem in implementing single assignment languages. In this paper, on the basis of a typical subset of the single assignment language SISAL, we analyze the inherent limits of storage reuse and define what the maximal storage reuse is. We propose an efficient method of achieving storage reuse, which is of polynomial complexity and linear in common cases. It achieves the maximal storage reuse for an extensive program class into which all common benchmark programs fall. We also show that no general method can guarantee the maximal storage reuse for programs outside the class. In this case, our method can choose the most likely operation to reuse the storage of an aggregate or a set of shared aggregates.

show abstract

Data flow graph optimization in if1

Cited by 29 publications

References 0 publications

Pythia: A Parallel Compiler for Delirium

Pythia: A Parallel Compiler for Delirium

Performance issues in dataflow machines

Efficient storage reuse of aggregates in single assignment languages

Contact Info

Product

Resources

About