Non-affine Extensions to Polyhedral Code Generation

Venkat, Anand; Shantharam, Manu; Hall, Mary; Strout, Michelle Mills

doi:10.1145/2581122.2544141

Cited by 44 publications

(31 citation statements)

References 31 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For a code to be raised to a polyhedral representation, all loop bounds, branch conditions and subscript functions must be affine expressions of outer loop iterators and constant parameters. Despite these constraints, many computational loops in scientific applications may be modeled through this representation either directly [12] or relying on existing model extensions [13], [14]. In our work, we consider only restrictions on the kernel loop bounds and on the subscript function of the monitored data as provided in the monitor construct (see Section II).…”

Section: A Backgroundmentioning

confidence: 99%

Adaptive Code Refinement: A Compiler Technique and Extensions to Generate Self-Tuning Applications

Schmitt

Helluy

Bastoul

2017

2017 IEEE 24th International Conference on High Performance Computing (HiPC)

View full text Add to dashboard Cite

Abstract-Compiler high-level automatic optimization and parallelization techniques are well suited for some classes of simulation or signal processing applications, however they usually don't take into account domain-specific knowledge nor the possibility to change or to remove some computations to achieve "good enough" results. Differently, production simulation and signal processing codes have adaptive capabilities: they are designed to compute precise results only where it matters if the complete problem is not tractable or if computation time must be short. In this paper, we present a new way to provide adaptive capabilities to compute-intensive codes automatically. It relies on domain-specific knowledge provided through special pragmas by the programmer in the input code and on polyhedral compilation techniques to continuously regenerate at runtime a code that performs heavy computations only where it matters. We present experimental results on several applications where our strategy enables significant computation savings and speedup while maintaining a good precision, with a minimal effort from the programmer.

show abstract

Section: A Backgroundmentioning

confidence: 99%

Adaptive Code Refinement: A Compiler Technique and Extensions to Generate Self-Tuning Applications

Schmitt

Helluy

Bastoul

2017

2017 IEEE 24th International Conference on High Performance Computing (HiPC)

View full text Add to dashboard Cite

show abstract

“…the non-zeros for each row should be adjacent in the data array) are not allowed in the SPF due to the restriction that all loop bounds are affine expressions of the surrounding loop iterators. This means that computations over sparse matrix data structures other than coordinate storage will need to be flattened with some form of loop restructuring [67].…”

Section: Composing a Legal Sequence Of Rtrtsmentioning

confidence: 99%

An approach for code generation in the Sparse Polyhedral Framework

et al. 2016

Self Cite

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

“…However, the polyhedral model is not applicable to (or intended for) one dimensional filter-like operations where the size of the result array depends on the source data. Recent work extends the polyhedral model to support arbitrary indexing [19], as well as conditional control flow that is predicated on arbitrary (ie, non-affine) functions of the loop indices [2]. However, the indices used to write into the destination array must still be computed with affine functions.…”

Section: Related Workmentioning

confidence: 99%

Fusing filters with integer linear programming

Robinson

Lippmeier

Keller

2014

Proceedings of the 3rd ACM SIGPLAN Workshop on Functional High-Performance Computing

View full text Add to dashboard Cite

The key to compiling functional, collection oriented array programs into efficient code is to minimise memory traffic. Simply fusing subsequent array operations into a single computation is not sufficient; we also need to cluster separate traversals of the same array into a single traversal. Previous work demonstrated how Integer Linear Programming (ILP) can be used to cluster the operators in a general data-flow graph into subgraphs, which can be individually fused. However, these approaches can only handle operations which preserve the size of the array, thereby missing out on some optimisation opportunities. This paper addresses this shortcoming by extending the ILP approach with support for size-changing operations, using an external ILP solver to find good clusterings.

show abstract

Non-affine Extensions to Polyhedral Code Generation

Cited by 44 publications

References 31 publications

Adaptive Code Refinement: A Compiler Technique and Extensions to Generate Self-Tuning Applications

Adaptive Code Refinement: A Compiler Technique and Extensions to Generate Self-Tuning Applications

An approach for code generation in the Sparse Polyhedral Framework

Fusing filters with integer linear programming

Contact Info

Product

Resources

About