SPPoC: manipulation automatique de polyèdres pour la compilation
作者： Pierre Boulet , Xavier Redon
DOI:
关键词:
摘要: The polyhedral model is quite popular in the field of parallelizing compilers. So, research prototypes tend to use tools like PIP (parametric integer programming solver), PolyLib (library for polyhedra manipulation) or Omega and calculator Presburger formulas). two main drawbacks these are a poor human-computer interface lack aggressive simplification. This last deficiency leads sequences computations which give too complex results even that cannot be completed due memory exhaustion time constraints. SPPoC (Symbolic Parameterized Polyhedral Calculator) library brings solution problems its completely symbolic advanced simpli- fication modules. It also allows unification different tools. We present applications : code generator communication volume estimator. MOTS-CLES Polyedres, Calcul symbolique, Generation de code, Transformations boucles, Data-parallelisme, Volume communications, PIP, PolyLib, Library
参考文章(14)
W. Kelly, W. Pugh, E. Rosser, Code generation for multiple mappings symposium on frontiers of massively parallel computation. pp. 332- 341 ,(1995) , 10.1109/FMPC.1995.380437
Michèle Dion, Yves Robert, Mapping affine loop nests: new results ieee international conference on high performance computing data and analytics. pp. 184- 189 ,(1995) , 10.1007/BFB0046628
Philippe Clauss, Counting solutions to linear and nonlinear constraints through Ehrhart polynomials: applications to analyze and transform scientific programs international conference on supercomputing. pp. 278- 285 ,(1996) , 10.1145/2591635.2667172
Paul Feautrier, Some efficient solutions to the affine scheduling problem: I. One-dimensional time International Journal of Parallel Programming. ,vol. 21, pp. 313- 348 ,(1992) , 10.1007/BF01407835
Paul Feautrier, Pierre Boulet, Scanning polyhedra without Do-loops international conference on parallel architectures and compilation techniques. pp. 4- 11 ,(1998) , 10.5555/522344.825714
A. Darte, Y. Robert, Affine-by-statement scheduling of uniform and affine loop nests over parametric domains Journal of Parallel and Distributed Computing. ,vol. 29, pp. 43- 59 ,(1995) , 10.1006/JPDC.1995.1105
Paul Feautrier, Dataflow analysis of array and scalar references International Journal of Parallel Programming. ,vol. 20, pp. 23- 53 ,(1991) , 10.1007/BF01407931
Michèle Dion, Yves Robert, Mapping affine loop nests parallel computing. ,vol. 22, pp. 1373- 1397 ,(1996) , 10.1016/S0167-8191(96)00049-X
William Pugh, The Omega test: a fast and practical integer programming algorithm for dependence analysis Proceedings of the 1991 ACM/IEEE conference on Supercomputing - Supercomputing '91. pp. 4- 13 ,(1991) , 10.1145/125826.125848
Philippe Clauss, Vincent Loechner, Parametric analysis of polyhedral iteration spaces application specific systems architectures and processors. ,vol. 19, pp. 179- 194 ,(1996) , 10.1023/A:1008069920230