Guest Editorial: Special Issue on Embedded Multicore Applications and Optimization
- 41 Downloads
Embedded systems with multi-core designs are becoming increasingly important for signal processing, machine learning, and multimedia applications from both industry and academia in recent years. While embedded multi-core systems will look to play an important role ahead for application designs, many challenging problems remain to be resolved. Applications, programming models, compilers, API designs, architecture designs, and software tools all need to contribute to the advance of embedded multi-core computing for signal processing, pixel processing, machine learning, and multimedia applications. In this special issue, we aim to bring together researchers in the related areas to present the latest developments and technical solutions concerning various aspects of embedded multi-core computing for signal processing and applications.
This special issue consists of four papers related to embedded multi-core computing. In the first paper, “A Halide-based Synergistic Computing Framework for Heterogeneous Systems ”, by Liao et al., is focused on the computing framework that extends Halide to improve program execution performance. Optimizations are developed to improve performance by generating adequate the parallel code and eliminating extra memory copies. The authors characterize and discuss the performance of two image processing programs and their framework on the heterogeneous platforms.
Second, the paper “GPUBlocks: GUI Programming Tool for CUDA and OpenCL ”, by Hwang et al., presents a GUI tool called GPUBlocks that can facilitate parallel programming on multicore computer systems. Programmers simply need to drag-n-drop blocks, fill the fields of the blocks, and connect them according to array or matrix computations that are specified by algorithms. GPUBlocks can then translate block-based code to CUDA or OpenCL programs. A couple of optimization constructs have also been offered for rapid program optimization.
For the third paper, “Rapid Hybrid Simulation for Exploring the Design Space of Signal Processors with Dynamic and Scalable Timing Models ”, authored by Yeh et al., presents a rapid hybrid emulation/simulation framework. This framework allows the user to execute full-blown system with application software and plug in emulators, simulators, and timing models for various components in the prototype system. In this system, with the just-in-time model selection mechanism it switches timing models dynamically.
The last paper in this special issue, entitled “Support OpenCL 2.0 Compiler on LLVM for PTX Simulators ” and authored by Yang et al., revises the low level virtual machine (LLVM) compiler to extend it to OpenCL 2.0 features to allow using the community GPGPU-Sim simulator.
The above four papers in this special issue cover language and simulation layers for embedded multi-core computing.. These papers provide frontier information related to heterogeneous computing, embedded compilers, and embedded multi-core programming tools.
The work is supported in part by research grants from Taiwan MOST and Mediatek.
- 1.Liao, S-W, Kuang, S-Y, Kao, C-L and Tu, C-H. (2019). A Halide-based Synergistic Computing Framework for Heterogeneous Systems, Journal of Signal Processing Systems, this issue.Google Scholar
- 2.Hwang, Y-S, Lin, H-H, Pai, S-H, and Tu, C-H. (2019). GPUBlocks: GUI Programming Tool for CUDA and OpenCL, Journal of Signal Processing Systems, this issue.Google Scholar
- 3.Yeh, C-W, Tu, C-H and Hung, S-H. (2019). Rapid Hybrid Simulation for Exploring the Design Space of Signal Processors with Dynamic and Scalable Timing Models, Journal of Signal Processing Systems, this issue.Google Scholar
- 4.Yang, C-C, Wang, S-C, Hsu, M-Y, Chang, Y-M, Hwang, Y-S and Lee, J-K. (2019). Support OpenCL 2.0 Compiler on LLVM for PTX Simulators, Journal of Signal Processing Systems, this issue.Google Scholar