TAIPEI, TAIWAN,Aug. 31, 2020 – Data is the foundation for all research in the scientific field and serves as a fundamental basis for intelligence. Therefore, the computation and acceleration of large amounts of data has become the focus for the development of high-performance computing. As the era of the internet and big data is emerging, the demand for data processing has subsequently been skyrocketing. However, as the computing of big data requires special programming consisting of a search and match analysis, the computing infrastructure will face challenges to cope with such massive amounts of data. While implementing large-scale hardware infrastructures can offer higher capacities to process big data, its tremendous financial costs can be a burden and limitation for its users.
Central Processing Units (CPUs) have taken the stage as the main computing power for the past 30 years; the increasing need for graphics and imaging computing has made Graphics processing units (GPUs) gradually become another main force for supercomputing in the past decade. However, there is a lack of another main computing force that specializes in processing non-image data. As the massive growth of data continues, imagining data after its pre-processing will become non-imaging data. As the cost for data storage is decreasing, internet processing speed is increasing; the consequent computing power needed is not keeping up with these trends. The lack of computing power shows a deviance in Moore’s Law for the development of semiconductors; therefore Moore’s law cannot be relied upon where computing power is concerned. Field Programmable Gate Arrays (FPGAs) are the best suited computing core to compensate for the ever-growing needs for the reduction of computing power. FPGAs also uphold itself to have the best computing and energy efficiency without sacrificing too much flexibility. Although some say that ASICs (Application Specific Integrated Circuits) are more efficient, the algorithms used by ASICs are fixed and not subjected to change as when the algorithm changes it would be invalid and obsolete. Therefore, compared to ASICs, FPGA has a high degree of flexibility and has an ability for quick modification; this allows the products to be quickly to the market. More importantly, FPGAs do not have high development costs. Therefore, IC design companies often use FPGA as the early stage of chip development as ASIC’s require a hefty initial investment cost.
Calvin, the CEO of WASAI Technology, mentioned, the advancement of technology brings improved computing efficiency and abilities; various processors have different advantages to meet the various needs of data centers. These processors can be explained with the analogy proposed by Calvin himself: the CPU is like a Swiss knife, the GPU is like a chainsaw, and the FPGA is like a carving knife for its precision and efficiency. We deem that FPGAs will be integrated into a large number of new computing applications in data centers within the next five years as we see a rapid development of edge computing data centers, 5G internet telecommunications, and the development of smart transportation and smart cities rising. As a large number of new computing applications in data centers will begin to integrate FPGAs, its appropriate computing architectures will be adjusted to meet different workloads to achieve data center computing to achieve maximum economic benefit.
About WASAI Technology Inc.
WASAI Technology's mission is to deliver acceleration technologies of High-Performance Data Analysis (HPDA) in future data centers for targeted vertical applications with massive volumes and high velocities of scientific data. To strengthen and advance scientific discovery and technological research via big data-intensive acceleration in high-performance computing, WASAI Technology aims to improve commercialization and commoditization of scientific and technological applications.
###
Comments