Artifact Evaluation for AT-Opt PACT 2018
<p>Artifact Evaluation Reproduction for "Optimizing Remote Data Transfers in X10", PACT 2018.</p> <p>------------------------------------------------------------------------------------------</p> <p>This repository contains artifacts and source codes<br> to reproduce experiments from the PACT 2018 research paper<br> titled "Optimizing Remote Data Transfers in X10"</p> <p>Our evaluations are done on Linux based Operation system.</p> <p><br> Hardware pre-requisities:<br> -------------------------<br> We recommend the following architectures any of the following architectures:<br> * An Intel System with two nodes, each node having 32 cores (32 CPUs). Total 64 cores.<br> * An AMD System, with two nodes, each node having 16 cores (16 CPUs). Total 32 cores.<br> </p> <p><br> Software pre-requisites:<br> ------------------------<br> 1) For Evaluation using Virtual Machine required software are already installed in the provided image.<br> A VirtualBox (from Oracle) is required to install the image and do evaluation.</p> <p>2) For Manual evalution, the below required software needs to be installed in the system.<br> * g++ (preferred version 5.4.0)<br> * Apache ant software (preferred version 1.9.6 )<br> * Java software<br> - preferred jvm jdk1.8.0_151 (or) java-8-oracle<br> - After installation, set JAVA_HOME path in ~/.bashrc to point to the installed JAVA bin</p> <p><br> Benchmarks:<br> -----------<br> 1) Taken from IMSuite Benchmark kernels (http://www.cse.iitm.ac.in/~krishna/imsuite/). Already, included in the package.</p> <p><br> Installation, Execution and Validation of results:<br> --------------------------------------------------<br> 1) For Evaluation using Virtual Machine see README.md file in ./AE/VM folder</p> <p>2) For Evaluation using Manual method see README.md file in ./AE/Manual folder</p> <p>3) For Evaluating AT-Opt technique againt varing input sizes (not discussed in the paper), look at<br> web page http://www.cse.iitm.ac.in/~krishna/imsuite/. In the ./AE/Manual folder, "x10-base" folder is<br> the baseline compiler (after bulid) and "x10-atOpt" (after build) folder is the AT-Opt compiler (has our techniques implemented). </p> <p><br> If anything in unclear, or any unexpected results occur, please report it to the authors</p>
ShareScore
32/100
Overall dataset sharing score
Score breakdown
These five areas show where the dataset supports — or may limit — practical reuse.
- Stewardship
- 4
- Harmonization
- 4
- Access
- 16
- Reuse readiness
- 8
- Engagement
- 0