Section: Highlights of the Year

Highlights of the Year

Aravind Sukumaran-Rajam has shown in his PhD work [13] that the polyhedral model, usually exclusively dedicated to advanced static analysis and optimization of linear loops, can also be applied to nonlinear loops. This noteworthy extension of the scope of polyhedral techniques has been made possible thanks to the speculative and dynamic parallelization strategy implemented in the Apollo framework. Significant parallel speed-ups can now be obtained automatically for loops and loop nest that could not be handled before by compilers. Aravind Sukumaran-Rajam and Philippe Clauss have published a paper on this topic in the ACM journal Transactions on Architecture and Code Optimization in 2015 [14] .