In the age of convergence between PCs and tablets, mobility and low power have become important metrics in judging a processor's value.
With the growing trends of green computing, BYOD and teleworking, form factors are shrinking, while expectations of battery life are increasing. Today's processors need to deliver robust performance across desktop-bound and mobile platforms with an ever-decreasing power envelope.
At the same time, workloads between PCs and mobile devices are increasingly similar, with compute, gaming, imaging and video tasks demanded from both mobile and desktop machines.
Adobe's Digital Index shows a 300 percent increase in video views from mobile devices in 2012, while PCs saw a 40 percent increase that cemented their lead as media consumption platforms. Additionally, in the enterprise, big data analytics is placing new demands on the processor, and platforms like Hadoop can benefit greatly from GPU acceleration.
As a result, the Central Processing Unit (CPU) and Graphics Processing Unit (GPU) are becoming increasingly intertwined. As I discussed in a recent interview, microprocessors are trending toward Accelerated Processing Units (APUs)-processors that combine CPU and GPU. 2014 will see the introduction of APUs with Heterogeneous System Architecture (HSA), an improved processor design that brings out the capabilities of different programmable compute elements, including the CPU and GPU, and makes them work together seamlessly to improve overall performance.
Impact on Asia
In Asia, these trends will affect a number of markets, catalyzing mobility adoption and lowering adoption costs for new technology.
Singapore has been described as an indicator of mobile trends and enjoys high cloud adoption. The growing role of tablets in collaborative design projects, document editing and project workspaces is driving demand for better performance from mobile processors. According to a Forrester report, tablets are expected to triple in sales by 2017 and reach majority status in Singapore and South Korea.
There are a number of growing markets in Asia as well. APUs are lowering the cost of ownership of a PC by negating the need for a discrete graphics card for computer-assisted design, digital content creation or multi-monitor setups. With smaller form factors, they are also addressing space constraints in workspaces. Therefore, despite the mature market, there is still untapped demand for PCs.
There has also been a shift toward GPU-based computing in the enterprise. Project Sumatra aims to enable Java developers to tap into the parallel processing power of the GPU to improve the performance of Java applications. IBM, for instance, has demonstrated that GPU-accelerated Java can boost the speed of data searching by 48 times-an application highly relevant to the growing big data industry.
There is also OpenCL, an open standard for writing programs that utilize CPUs, GPUs and other processors. OpenCL is supported by projects like clMath, a set of OpenCL libraries designed to harness APUs and GPUs. Industries that demand high amounts of parallel computing power, like advanced imaging, cryptography and bioinformatics, can take advantage of OpenCL to gain greater performance from their hardware.
Sign up for Computerworld eNewsletters.