Loading...
Please wait, while we are loading the content...
Similar Documents
Writing a modular gpgpu program in java.
| Content Provider | CiteSeerX |
|---|---|
| Author | Ioki, Masayuki Hozumi, Shumpei Chiba, Shigeru |
| Abstract | www.csg.is.titech.ac.jp This paper proposes a Java to CUDA runtime program trans-lator for scientific-computing applications. Traditionally, these applications have been written in Fortran or C with-out using a rich modularization mechanism. Our translator enables those applications to be written in Java and run on GPGPUs while exploiting a rich modularization mechanism in Java. This translator dynamically generates optimized CUDA code from a Java program given at bytecode level when the program is running. By exploiting dynamic type information given at translation, the translator devirtualizes dynamic method dispatches and flattens objects into simple data representation in CUDA. To do this, a Java program must be written to satisfy certain constraints. This paper also shows that the performance overheads due to Java and WootinJ are not significantly high. |
| File Format | |
| Access Restriction | Open |
| Subject Keyword | Modular Gpgpu Program Java Program Rich Modularization Mechanism Bytecode Level Dynamic Type Information Certain Constraint Cuda Runtime Program Trans-lator Flattens Object Performance Overhead Optimized Cuda Code Dynamic Method Dispatch Scientific-computing Application Simple Data Representation |
| Content Type | Text |
| Resource Type | Article |
