Single Assignment C (SAC) High Productivity Meets High Performance

Grelck, Clemens

doi:10.1007/978-3-642-32096-5_5

Cited by 14 publications

(8 citation statements)

References 50 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…As the name suggests, SaC leaves the beaten track of functional languages and adopts a C-like syntax: we interpret assignment sequences as nested let-expressions, branching constructs as conditional expressions and loops as syntactic sugar for tail-recursive functions. For any details regarding language design we refer the interested reader to [1,2].…”

Section: Sac-single Assignment Cmentioning

confidence: 99%

“…Note that the statement A_old = A; does not copy the array, as for instance in superficially similar Fortran-90 code, but merely creates a new lambda-binding to an existing (array) value. For a complete discussion of the example we refer the interested reader to [8], for a thorough introduction of SaC to [2].…”

Section: Sac-single Assignment Cmentioning

confidence: 99%

“…For thorough discussions of our first benchmark regarding both general language design and asynchronous adaptive specialization we refer the interested reader to [2,8], respectively. In essence, the benchmark alternately computes one convolution step and the convergence check.…”

Section: Rank-generic Convolution With Convergence Testmentioning

confidence: 99%

“…Such scenarios motivate our current research in the context of the purely functional, data-parallel array language SaC (Single Assignment C) [1,2]. SaC features immutable, homogeneous, multi-dimensional arrays and supports both shape-and rank-generic programming: SaC functions may not only abstract from the concrete shapes of argument and result arrays, but likewise from their ranks (i.e.…”

Section: Introductionmentioning

confidence: 99%

See 3 more Smart Citations

Persistent Asynchronous Adaptive Specialization for Generic Array Programming

Grelck

Wiesinger

2018

Int J Parallel Prog

Self Cite

View full text Add to dashboard Cite

Generic array programming systematically abstracts from structural array properties such as shape and rank. As usual, generic programming comes at the price of lower runtime performance. The idea of asynchronous adaptive specialization is to exploit parallel computing facilities to reconcile these conflicting objectives through the continuous adaptation of running applications to the ranks and shapes of their arrays. A key parameter for the effectiveness of our approach is the time it takes from requesting a certain specialization until its availability to the running application. We describe the ins and outs of a persistence layer that keeps specialized variants in a repository for future use and thus effectively reduces the average waiting time for re-compilation to nearly zero. A number of critical issues that, among others, stem from the interplay between function specialization and function overloading catch our special attention. We describe the solutions adopted and illustrate the benefits of persistent asynchronous adaptive specialization by a series of experiments.

show abstract

Section: Sac-single Assignment Cmentioning

confidence: 99%

Section: Sac-single Assignment Cmentioning

confidence: 99%

Section: Rank-generic Convolution With Convergence Testmentioning

confidence: 99%

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Persistent Asynchronous Adaptive Specialization for Generic Array Programming

Grelck

Wiesinger

2018

Int J Parallel Prog

Self Cite

View full text Add to dashboard Cite

show abstract

“…On recalls that, for functional imperative parallel programming, algorithm skeleton [6,9] has been a popular approach to composing parallel constructs such as divide-and-conquer, master-slave, pipe, and so on. Implicit data-parallel languages, such as SAC [10], have long existed for specifying array-based computation. Cellular automaton operations can also be specified over regular geometric grids [11].…”

mentioning

confidence: 99%

GridFOR: A Domain Specific Language for Parallel Grid-Based Applications

Wang

2015

Int J Parallel Prog

View full text Add to dashboard Cite

To ease the programming burden and to make parallel programs more maintainable, computational scientists and engineers currently have the options to use software libraries, templates, and general purpose language extensions to compose their application programs. These existing options, unfortunately, have considerable limitations with compatibility, expressive power and delivered performance. To address these issues, we design a domain specific language, GridFOR, for computational problems defined over regular geometric grids. This language allows the programmer to first implement an algorithm on simple data structures, as commonly illustrated in textbooks or papers. The programmer then specifies transformations to extend the algorithm for complex data structures required by the target applications. We build a compiler to automatically translate a GridFOR program to a parallel Fortran version with Message Passing Interface calls. Several optimization techniques are implemented in our compiler to enhance the program speed.

show abstract

SaC/C formulations of the all‐pairs N‐body problem and their performance on SMPs and GPGPUs

Šinkarovs

Scholz

Bernecky³

et al. 2013

Concurrency and Computation

Self Cite

View full text Add to dashboard Cite

This paper describes our experience in implementing the classical N -body algorithm in SAC and analysing the runtime performance achieved on three different machines: a dual-processor 8-core Dell PowerEdge 2950 (a Beowulf cluster node, the reference machine), a quad-core hyper-threaded Intel Core-i7 based system equipped with an NVidia GTX-480 graphics accelerator and an Oracle Sparc T4-4 server with a total of 256 hardware threads. We contrast our findings with those resulting from the reference C code and a few variants of it that employ OpenMP pragmas as well as explicit vectorisation. Our experiments demonstrate that the SAC implementation successfully combines a high level of abstraction, very close to the mathematical specification, with very competitive runtimes. In fact, SAC matches or outperforms the hand-vectorised and hand-parallelised C codes on all three systems under investigation without the need for any source code modification. Furthermore, only SAC is able to effectively harness the advanced compute power of the graphics accelerator, again by mere recompilation of the same source code. Our results illustrate the benefits that SAC provides to application programmers in terms of coding productivity, source code, and performance portability among different machine architectures, as well as long-term maintainability in evolving hardware environments.

show abstract

Single Assignment C (SAC) High Productivity Meets High Performance

Cited by 14 publications

References 50 publications

Persistent Asynchronous Adaptive Specialization for Generic Array Programming

Persistent Asynchronous Adaptive Specialization for Generic Array Programming

GridFOR: A Domain Specific Language for Parallel Grid-Based Applications

SaC/C formulations of the all‐pairs N‐body problem and their performance on SMPs and GPGPUs

Contact Info

Product

Resources

About