The main goal of the RAPID framework is to allow low-power devices to support compute-intensive tasks and applications at reasonable time. Hence, RAPID aspires to enable devices such as smartphones, notebooks, tablets, wearable devices, etc. to deliver both Central Processing Unit (CPU) and GPU intensive tasks, exploiting resources of more capable devices, or even the cloud, participating in the RAPID ecosystem. In this perspective, RAPID is developing an efficient heterogeneous CPU-GPU cloud computing infrastructure, which can be used to seamlessly offload CPU-based and GPU-based (using CUDA) tasks of applications to more powerful devices over a heterogeneous network (HetNet).
This library is part of the RAPID Project and is an ongoing work.
While RAPID envisions to support heterogeneous devices, this is the component that enables code offloading on Android.
We have provided some initial demo applications together with instructions on how to use the library.
Check them on the RAPID Android Demo repository.
For Java and CUDA code offloading on other platforms, have a look at the generic RAPID Linux/Windows Framework.