Heterogeneous computing: challenges and opportunities

Abstract

Anytime you work with oranges and apples, you'll need a number of schemes to organize total performance. This article surveys the challenges posed by H omogeneous computing, which uses one or more machines of the same type, has provided adequate performance for many applications in the past. Many of these applications had more than one type of embedded parallelism, such as single instruction, multiple data (SIMD) and multiple instruction, multiple data (MIMD). Most of the current parallel machines are suited only for homogeneous computing. However, numerous applications that have more than one type of embedded parallelism are now being considered for parallel implementation. On the other hand, as the amount of homogeneous parallelism in applications decreases, homogeneous systems cannot offer the desired speedups. To exploit the heterogeneity in computations, researchers are investigating a suite of heterogeneous architectures.Heterogeneous computing (HC) is the well-orchestrated and coordinated effective use of a suite of diverse high-performance machines (including parallel machines) to provide superspeed processing for computationally demanding tasks with diverse computing needs.' An HC system includes heterogeneous machines, high-speed networks, interfaces, operating systems, communication protocols, and programming environments, all combining to produce a positive impact on ease of use and performance. Figure 1 shows an example HC environment.Heterogeneous computing should be distinguished from network computing or high-performance distributed computing, which have generally come to mean either clusters of workstations or ad hoc connectivity among computers using little more than opportunistic load-balancing. HC is a plausible, novel technique for heterogeneous solving computationally intensive problems that have several types of embedded computing and parallelism. HC also helps to reduce design risks by incorporating proven technology and existing designs instead of developing them from scratch. However, discusses some several issues and problems arise from employing this technique, which we discuss.In the past few years, several technical meetings have addressed many of these approaches to opening issues. There is also a growing interest in using this paradigm to solve Grand up its opportunities.

A Static Task Partitioning Approach for Heterogeneous Systems Using OpenCL

“…By having processing units with different characteristics, computation can be mapped to specialised devices that perform a specific type of task more efficiently than other devices. In embedded systems this has been the case for many years with specialised DSP units for instance [15]. This trend has spread to the desktop, where the high-end relies on accelerator devices for increased performance.…”

“…Heterogeneous computing systems promise to deliver high performance at relatively low energy costs [15,18]. By having processing units with different characteristics, computation can be mapped to specialised devices that perform a specific type of task more efficiently than other devices.…”

Dynamically mapping tasks with priorities and multiple deadlines in a heterogeneous environment

“…Determination of ETC values is a separate research problem, and the assumption of such ETC information is a common practice in mapping research (e.g., [16,20,21,24,31,39]). …”

“…The estimated time to complete (ETC) values of each task on each machine is assumed to be known based on user-supplied information, experiential data, task profiling and analytical benchmarking, or other techniques (e.g., [1,15,16,21,26,40]). Determination of ETC values is a separate research problem, and the assumption of such ETC information is a common practice in mapping research (e.g., [16,20,21,24,31,39]).…”

Task mapping on supercomputers with cellular networks

“…Heterogeneous computing (HC) is an efficient technique to solve computationally intensive problems that has several types of embedded parallelism [9]. The problem of mapping large and diverse groups of tasks onto the machines of an HC had been researched by Siegel et al The mapping problem is formulated as follows [79]:…”