IDC Analyzes Accelerated Computing through Digital Innovation Technologies Including Cognitive Computing
Expected to Impact Virtually All Workloads

According to recent research and analysis by IDC (www.idc.com), the position of accelerated computing is strengthening to overcome the limitations of conventional CPUs, and adoption by enterprises is increasing.

IDC recently announced a classification system for accelerated computing to help enterprises understand the field that accelerated computing occupies in the computing platform sector and to support deployment strategies.

Accelerated computing is a function that accelerates applications and workloads by allocating portions of processes to silicon subsystems such as graphics processing units (GPUs) or FPGAs. It is gaining attention from enterprises seeking to overcome the limitations of central processing units (CPUs) in workloads that require data processing acceleration.

Accelerated computing is used in cognitive computing, deep learning, artificial intelligence, machine learning and similar types of applications, data analytics workloads including visual analytics, scientific and technical workloads, cloud computing and acceleration as a service, and unstructured data management workloads including edge computing. Accelerated computing is expected to impact virtually all workloads.

Peter Rutten, Senior Research Director at IDC's Server and Computing Platform research division, stated, "As current acceleration technologies such as GPUs and FPGAs begin to transform server infrastructure to meet workload performance requirements including cognitive and AI applications, future computing will look different from today," and explained that the fields in which accelerated computing will be utilized are limitless due to the capabilities and technical characteristics of acceleration devices that can be deployed to suit workloads.

According to IDC's recent survey results, when asked what attributes are important in deploying infrastructure related to mission-critical workloads within enterprises, approximately three-quarters of respondents answered single or multiple GPUs. GPUs are particularly attractive to enterprises because they use standard-type libraries that can be easily integrated with applications and can be purchased in commodity form. However, other technologies that potentially provide higher performance per watt, such as FPGAs, multi-core processors, and application-specific integrated circuits (ASICs), are also beginning to receive attention.

Kwon Sang-jun, Senior Research Director at Korea IDC, stated, "Robotics, cognitive systems, IoT, virtual reality, and augmented reality, which are drawing attention in digital transformation, require the swift and accurate analysis of vast amounts of unstructured data to derive optimal results above all else," adding, "Accelerated computing is a key technology that can drive digital innovation and is expected to become an essential element in strengthening enterprise capabilities."