A Hybrid Approach to Proving Memory Reference Monotonicity

作者: Cosmin E. Oancea , Lawrence Rauchwerger

DOI: 10.1007/978-3-642-36036-7_5

关键词:

摘要: Array references indexed by non-linear expressions or subscript arrays represent a major obstacle to compiler analysis and automatic parallelization. Most previous proposed solutions either enhance the static repertoire recognize more patterns, infer array-value properties, refine mathematical support, apply expensive run time of memory reference traces disambiguate these accesses. This paper presents an automated solution based on construction access summaries, in which non-linearity problem can be solved for large number patterns extracting arbitrarily-shaped predicates that (in)validate monotonicity property thus (dis)prove loop independence. Experiments six benchmarks show our general technique dynamic validation cover class codes, incurs minimal run-time overhead obtains good speedups.

参考文章(28)
Silvius Rus, Lawrence Rauchwerger, Jay Hoeflinger, Hybrid analysis: static & dynamic memory reference analysis International Journal of Parallel Programming. ,vol. 31, pp. 251- 283 ,(2003) , 10.1023/A:1024597010150
Cosmin E. Oancea, Alan Mycroft, Set-Congruence Dynamic Analysis for Thread-Level Speculation (TLS) languages and compilers for parallel computing. pp. 156- 171 ,(2008) , 10.1007/978-3-540-89740-8_11
Yuan Lin, David Padua, Demand-Driven Interprocedural Array Property Analysis languages and compilers for parallel computing. pp. 303- 317 ,(1999) , 10.1007/3-540-44905-1_19
Unified Interprocedural Parallelism Detection International Journal of Parallel Programming. ,vol. 29, pp. 185- 215 ,(2001) , 10.1023/A:1007685003043
Thomas Fahringer, Efficient Symbolic Analysis for Parallelizing Compilers and Performance Estimators The Journal of Supercomputing. ,vol. 12, pp. 227- 252 ,(1998) , 10.1023/A:1008013906463
William Pugh, David Wonnacott, Nonlinear array dependence analysis Springer, Boston, MA. pp. 1- 14 ,(1994) , 10.1007/978-1-4615-2315-4_1
Sungdo Moon, Mary W. Hall, Evaluation of predicated array data-flow analysis for automatic parallelization acm sigplan symposium on principles and practice of parallel programming. ,vol. 34, pp. 84- 95 ,(1999) , 10.1145/301104.301112
Yunheung Paek, Jay Hoeflinger, David Padua, Efficient and precise array access analysis ACM Transactions on Programming Languages and Systems. ,vol. 24, pp. 65- 109 ,(2002) , 10.1145/509705.509708