Automatic Parallelization using AutoFutures
- 
Name:Konferenzartikel 
- 
Author:Korbinian Molitorisz, Jochen Schimmel, Frank Otto 
- 
                    ZusammenfassungPractically all new computer systems are parallel. The mindsof the majority of software engineers are not, and most of existing sourcecode is still sequential. Within only a few years, multicore processorschanged the system landscape, but the competence to reengineer forcomputer systems of today is shared among a small community of soft-ware engineers.In this paper we present AutoFuture, an approach that automaticallyidentifies parallelizable locations in sequential source code and reengi-neers them for multicore. This approach demands minimal change tosequential source code. AutoFutures make parallel code easy to under-stand and increase the acceptance of parallel software. 
- 
Year:2012 
- Links:
| Titel Vorname Nachname | 
|---|
| Dipl.-Inform. Jochen Schimmel | 
| Dr. Ing. Korbinian Molitorisz | 
| Dr. Frank Otto | 
| Titel | 
|---|
| Forschungsprojekt AParT - Entwurfsmustergestütze Anwendungsparallelisierung | 
| SRG Entwurfsmustergestützte Anwendungsparallelisierung | 
Bibtex
@inproceedings{,
author={Korbinian Molitorisz, Jochen Schimmel, Frank Otto},
title={Automatic Parallelization using AutoFutures},
year=2012,
month=May,
booktitle={International Conference on Multicore Software Engineering, Performance, and Tools (MSEPT'12)},
url={https://ps.ipd.kit.edu/donloads/},
abstract={Practically all new computer systems are parallel. The mindsof the majority of software engineers are not, and most of existing sourcecode is still sequential. Within only a few years, multicore processorschanged the system landscape, but the competence to reengineer forcomputer systems of today is shared among a small community of soft-ware engineers.In this paper we present AutoFuture, an approach that automaticallyidentifies parallelizable locations in sequential source code and reengi-neers them for multicore. This approach demands minimal change tosequential source code. AutoFutures make parallel code easy to under-stand and increase the acceptance of parallel software.},
journal={International Conference on Multicore Software Engineering, Performance, and Tools (MSEPT'12)},
} 