International audienceArray contraction is a compilation optimization used to reduce memory consumption, by reducing the size of temporary arrays in a program while preserving its correctness. The usual approach to this problem is to perform a static analysis of the given program, creating overhead in the compilation cycle. In this work, we take a look at exploiting execution traces of programs of the polyhedral model, in order to infer reduced sizes for the temporary arrays used during calculations. We designed a four step process to reduce the storage requirements of a temporary array of a given scheduled program, in which we used an algorithm to deduce array access functions for which bounds are modulos of affine functions of parameters ...
International audienceThe ARGO H2020 European project aims at developing a Worst-Case Execution Time...
This is a post-peer-review, pre-copyedit version of an article published in IEEE Transactions on Com...
The polyhedron model has been used successfully for auto-matic parallelization of code regions with ...
International audienceArray contraction is a compilation optimization used to reduce memory consumpt...
Array contraction is a compilation optimization used to reduce the memory con-sumption, b...
International audienceMemory accesses are a well known bottleneck whose impact might be mitigated by...
Memory accesses are a well known bottleneck whose impact might be mitigated by using properly the me...
The polyhedral model has been successfully used in production compilers. Nevertheless, only a very r...
International audienceWe build on prior work on intra-array memory reuse, for which a general theore...
Many modern (mobile) systems involve memory intensive computations. External memory accesses are cos...
International audienceWe study the parallelizing compilation and loop nest optimization of an import...
Multi-core processors are now in widespread use in almost all areas of computing: desktops, laptops ...
International audienceIncreasing data movement costs motivate the integration of polyhedral loop opt...
A formal, high-level representation of programs is typically needed for static and dynamic analyses ...
Efficient memory usage is crucial for data-intensive applications as a smaller memory footprint ensu...
International audienceThe ARGO H2020 European project aims at developing a Worst-Case Execution Time...
This is a post-peer-review, pre-copyedit version of an article published in IEEE Transactions on Com...
The polyhedron model has been used successfully for auto-matic parallelization of code regions with ...
International audienceArray contraction is a compilation optimization used to reduce memory consumpt...
Array contraction is a compilation optimization used to reduce the memory con-sumption, b...
International audienceMemory accesses are a well known bottleneck whose impact might be mitigated by...
Memory accesses are a well known bottleneck whose impact might be mitigated by using properly the me...
The polyhedral model has been successfully used in production compilers. Nevertheless, only a very r...
International audienceWe build on prior work on intra-array memory reuse, for which a general theore...
Many modern (mobile) systems involve memory intensive computations. External memory accesses are cos...
International audienceWe study the parallelizing compilation and loop nest optimization of an import...
Multi-core processors are now in widespread use in almost all areas of computing: desktops, laptops ...
International audienceIncreasing data movement costs motivate the integration of polyhedral loop opt...
A formal, high-level representation of programs is typically needed for static and dynamic analyses ...
Efficient memory usage is crucial for data-intensive applications as a smaller memory footprint ensu...
International audienceThe ARGO H2020 European project aims at developing a Worst-Case Execution Time...
This is a post-peer-review, pre-copyedit version of an article published in IEEE Transactions on Com...
The polyhedron model has been used successfully for auto-matic parallelization of code regions with ...