A Hierarchical Model of Reference Affinity

Structure Data Layout Optimization (SDLO) is a prevailing compiler optimization technique to improve cache efficiency. Structure transformation is a critical step for SDLO. Diversity of transformation methods and existence of complex data types are major challenges for structure transformation. We have designed and implemented STrans, a well-defined system which provides controllable and comprehensive functionality on structure transformation. Compared to known systems, it has less limitation on data types for transformation.In this paper we give formal definition of the approach STrans transforms data types. We have also designed Transformation Specification Language, a mini language to configure how to transform structures, which can be either manually tuned or generated by compiler. STrans supports three kinds of transformation methods, i.e., splitting, peeling, and pool-splitting, and works well on different combinations of compound data types. STrans is the transformation system used in ASLOP and is well tested for all benchmarks for ASLOP.

show abstract

“…There are many papers focusing on making decision on a good SDLO or general data layout strategy [2,3,[12][13][14][15][16] .…”

Section: Related Workmentioning

confidence: 99%

STrans: a comprehensive framework for structure transformation

Chen

Tang

2016

Tinshhua Sci. Technol.

View full text Add to dashboard Cite

show abstract

“…Zhong et al [16] defines a hierarchical model, called reference affinity, based on which, their further work [2] proposes k-distance analysis that are used to guide structure data-layout. Reference affinity is one way to measure affinity of different fields in the same instance.…”

Section: Related Workmentioning

confidence: 99%

ASLOP: A field-access affinity-based structure data layout optimizer

Yan

Chen

et al. 2011

Sci. China Inf. Sci.

View full text Add to dashboard Cite

By rearranging the data, data layout optimizations improve the utilization of a cache line between two of its successive refills, thus reducing the total number of cache line refills and improving the performance of a program. In this paper, we show that to enable structure data layout optimizations to be effective, two parameters, namely intra-instance affinity and inter-instance affinity, need to be considered at the same time in order to model the cache line utilization more accurately. We also propose a lightweight approach to measure intra-instance affinity and inter-instance affinity to avoid complex memory trace analyses. A prototype, called ASLOP, has been implemented in the Open64 compiler and evaluated using benchmarks from SPEC CPU 2000, SPEC CPU 2006 and Olden benchmark suites that have extensive structure types. Our approach can achieve up to 48.1% performance improvement over the original programs, and 11.9% over the optimized programs using maximal reshaping, an existing approach that is known to produce close to the best results, on the two platforms we tested.

show abstract

“…A different approach is to restructure traces to yield new insights. Reference affinity [31] is one such example that places a hierarchical ordering on a reference trace, grouping together correlated references as much as possible, generalizing the notion of spatial and temporal locality. The structure can be used as a guide to select caching strategies, or more generally to understand the access pattern present in a trace.…”

Section: Related Workmentioning

confidence: 99%

Topological analysis and visualization of cyclical behavior in memory reference traces

Choudhury

Wang

Rosen

et al. 2012

2012 IEEE Pacific Visualization Symposium

View full text Add to dashboard Cite

(a) (b) Figure 1: One circular parameterization of the memory behavior of bubble sort (box (a)) morphing into another (box (b)), highlighting both their similarities and differences, and giving two views of the recurrent nature of the program. ABSTRACTWe demonstrate the application of topological analysis techniques to the rather unexpected domain of software visualization. We collect a memory reference trace from a running program, recasting the linear flow of trace records as a high-dimensional point cloud in a metric space. We use topological persistence to automatically detect significant circular structures in the point cloud, which represent recurrent or cyclical runtime program behaviors. We visualize such recurrences using radial plots to display their time evolution, offering multi-scale visual insights, and detecting potential candidates for memory performance optimization. We then present several case studies to demonstrate some key insights obtained using our techniques.

show abstract

A Hierarchical Model of Reference Affinity

Cited by 3 publications

References 36 publications

STrans: a comprehensive framework for structure transformation

STrans: a comprehensive framework for structure transformation

ASLOP: A field-access affinity-based structure data layout optimizer

Topological analysis and visualization of cyclical behavior in memory reference traces

Contact Info

Product

Resources

About