europar2020.zip (10.65 MB)
Artifact and instructions to generate experimental results for Euro-Par proceeding 2020 paper: Skipping Non-essential Instructions Makes Data-dependence Profiling Faster
software
posted on 2020-07-03, 15:05 authored by Nicolas Morew, Mohammad Norouzi, Ali Jannesari, Felix WolfThis artifact is concerned with Section 4 (Evaluation) of
the Paper "Skipping Non-essential Instructions Makes Data-dependence Profiling Faster" by Nicolas Morew, Mohammad Norouzi, Ali Jannesari and Felix Wolf. The artifact consists of
the llvm pass to build binaries of the two data-dependence profilers: DiscoPoP and the hybrid approach which is presented in Section 3 of the paper.
Also, it contains the benchmarks which were used to produce the results presented in the paper. Please find a description of the artifact in the Overview document (word and .pdf formats are provided).
This artifact is provided as a single .zip archive. Scripts are provided in bash shell .sh format and .py format. Additional copyright, source and license details are provided in the 'license' file.
the Paper "Skipping Non-essential Instructions Makes Data-dependence Profiling Faster" by Nicolas Morew, Mohammad Norouzi, Ali Jannesari and Felix Wolf. The artifact consists of
the llvm pass to build binaries of the two data-dependence profilers: DiscoPoP and the hybrid approach which is presented in Section 3 of the paper.
Also, it contains the benchmarks which were used to produce the results presented in the paper. Please find a description of the artifact in the Overview document (word and .pdf formats are provided).
This artifact is provided as a single .zip archive. Scripts are provided in bash shell .sh format and .py format. Additional copyright, source and license details are provided in the 'license' file.