Abstract
The paper introduces ten non-serial polyadic dynamic programming (NPDP) kernels as a benchmark suite dedicated to effectiveness evaluation of tiled code generated by means of polyhedral optimization compilers. Most of the applications implement bioinformatics algorithms which are challenging and ongoing tasks for automatic loop nest tiling transformations. The paper describes mathematically examined kernels and uniformizes them in the form of loop nests presented in the C language. In an experimental study, we applied the two automatic source-to-source compilers, TRACO and PluTo, to generate cache-efficient codes and analysed their performance on three multi-core machines. We discuss the limitations of well-known tiling approaches and outline future tiling strategies for the introduced NPDP Benchmark suite.
Sign-in/Register to access full text options 
Free) 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call
