Abstract
Current HPC clusters are composed by several machines with different computation capabilities and different kinds and families of accelerators. Programming efficiently for these heterogeneous systems has become an important challenge. There are many proposals to simplify the programming and management of accelerator devices, and the hybrid programming, mixing accelerators and CPU cores. However, in many cases, portability compromises the efficiency on different devices, and there are details concerning the coordination of different types of devices that should still be tackled by the programmer. In this work, we introduce the Multi-Controller, an abstract entity implemented in a library that coordinates the management of heterogeneous devices, including accelerators with different capabilities and sets of CPU-cores. Our proposal improves state-of-the-art solutions, simplifying data partition, mapping and the transparent deployment of both, simple generic kernels portable across different device types, and specialized implementations defined and optimized using specific native or vendor programming models (such as CUDA for NVIDIA’s GPUs, or OpenMP for CPU-cores). The run-time system automatically selects and deploys the most appropriate implementation of each kernel for each device, managing data movements and hiding the launch details. The results of an experimental study with five study cases indicates that our abstraction allows the development of flexible and highly efficient programs that adapt to the heterogeneous environment.
Highlights
The Multi-Controller library we are presenting in this paper is built on top of two previous tools
In this work we introduce the Multi-Controler (MCtrl), an abstract entity implemented in a library, that coordinates the management of heterogeneous devices, including accelerators with different capabilities and sets of CPUcores
Our proposal improves state-of-the-art solutions, simplifying the data partition, mapping, and transparent deployment of both, simple generic kernels portable across different device types, and specialized implementations defined and optimized using specific native or vendor programming models
Summary
The Multi-Controller library we are presenting in this paper is built on top of two previous tools. The first one, named Hitmap [3], is a library to manage the partition and mapping of data structures. It is used in our model to manage the data distribution across devices, and to provide a common interface to implement data management inside generic portable kernels. The second one, named Controller [1, 12], is a library that defines an abstract entity to transparently manage the data movements and kernel launching for a single device. Our proposal combines them with a new layer of abstraction to coordinate the use of several devices of different architectures or natures. This section introduces the reader to both previous tools, before describing the new abstraction proposed
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
