Runtime Code Generation in C++ as a Foundation for Domain-Specific Optimisation

Beckmann, Olav; Houghton, Alastair; Mellor, Michael R.

doi:10.1007/978-3-540-25935-0_17

Cited by 35 publications

(24 citation statements)

References 19 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Such static analyses should also enable the design of smarter (higherlevel) transformation primitives. In the longer term, we also wish to invest in a more robust implementation of the X language, based on a run-time compilation framework, like ROSE [22] or TaskGraph [1], and/or using a more abstract code representation in the polytope model [5]. Our main long-term goal is the adoption by application experts with little interest in compiler design and implementation.…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

A Language for the Compact Representation of Multiple Program Versions

Donadio

Brodman

Roeder

et al. 2006

Languages and Compilers for Parallel Computing

View full text Add to dashboard Cite

Abstract. As processor complexity increases compilers tend to deliver suboptimal performance. Library generators such as ATLAS, FFTW and SPIRAL overcome this issue by empirically searching in the space of possible program versions for the one that performs the best. Empirical search can also be applied by programmers, but because they lack a tool to automate the process, programmers need to manually re-write the application in terms of several parameters whose best value will be determined by the empirical search in the target machine. In this paper, we present the design of an annotation language, meant to be used either as an intermediate representation within library generators or directly by the programmer. This language that we call X represents parameterized programs in a compact and natural way. It provides an powerful optimization framework for high performance computing.

show abstract

Section: Discussionmentioning

confidence: 99%

“…The TaskGraph library [1] is closely related with the X language. It combines a structured multistage evaluation layer built on top of C++ expression templates, with run-time generation and compilation, and with a transformation toolkit based on SUIF (1.3) [11] and/or ROSE [22].…”

Section: Related Workmentioning

confidence: 99%

A Language for the Compact Representation of Multiple Program Versions

Donadio

Brodman

Roeder

et al. 2006

Languages and Compilers for Parallel Computing

View full text Add to dashboard Cite

show abstract

“…This distinguishes POET from the large, existing body of work on powerful languages and tools for expressing static code transformations [20], [1]. We intend to use POET in the context of an empirical search process; we do not specifically address run-time code generation as performed by more general multistage languages and systems [2], [7], [10], [11].…”

Section: Related Workmentioning

confidence: 99%

POET: Parameterized Optimizations for Empirical Tuning

Seymour

You

et al. 2007

2007 IEEE International Parallel and Distributed Processing Symposium

View full text Add to dashboard Cite

show abstract

“…In previous work [5], we defined criteria for what we call language virtualization, saying that a general-purpose language is virtualizable iff it can provide an environment to embedded languages that makes them essentially identical to corresponding stand-alone language implementations in terms of expressiveness (being able to express a DSL in a way which is natural to domain specialists) performance (leveraging domain knowledge to produce optimal code), and safety (domain programs are guaranteed to have certain properties implied by the DSL), while at the same time requiring only modestly more development effort than implementing a simple, pure-library embedding. 1 One ingredient of LMS is a finally tagless [4] or polymorphic [20] language embedding, which ensures expressiveness and safety. Hofer et al [20] show that a polymorphic embedding can be constructed from a pure embedding [21] with acceptable effort.…”

Section: Introductionmentioning

confidence: 99%

Lightweight modular staging

RompfTiark

OderskyMartin

2010

SIGPLAN Not.

View full text Add to dashboard Cite

Software engineering demands generality and abstraction, performance demands specialization and concretization. Generative programming can provide both, but developing high-quality program generators takes a large effort, even if a multi-stage programming language is used.We present lightweight modular staging, a library-based multistage programming approach that breaks with the tradition of syntactic quasi-quotation and instead uses only types to distinguish between binding times. Through extensive use of component technology, lightweight modular staging makes an optimizing compiler framework available at the library level, allowing programmers to tightly integrate domain-specific abstractions and optimizations into the generation process.We argue that lightweight modular staging enables a form of language virtualization, i.e. allows to go from a pure-library embedded language to one that is practically equivalent to a standalone implementation with only modest effort.

show abstract

Runtime Code Generation in C++ as a Foundation for Domain-Specific Optimisation

Cited by 35 publications

References 19 publications

A Language for the Compact Representation of Multiple Program Versions

A Language for the Compact Representation of Multiple Program Versions

POET: Parameterized Optimizations for Empirical Tuning

Lightweight modular staging

Contact Info

Product

Resources

About