The Paramountcy of Reconfigurable Computing [chapter]

Reiner Hartenstein
2012 Energy-Efficient Distributed Computing Systems  
Computers are very important for all of us. But brute force disruptive architectural developments in industry and threatening unaffordable operation cost by excessive power consumption are a massive future survival problem for our existing cyber infrastructures, which we must not surrender. The progress of performance in high performance computing (HPC) has stalled because of the "programming wall" caused by lacking scalability of parallelism. This chapter shows that Reconfigurable Computing is
more » ... the silver bullet to obtain massively better energy efficiency as well as much better performance, also by the upcoming methodology of HPRC (high performance reconfigurable computing). We need a massive campaign for migration of software over to configware. Also because of the multicore parallelism dilemma, we anyway need to redefine programmer education. The impact is a fascinating challenge to reach new horizons of research in computer science. We need a new generation of talented innovative scientists and engineers to start the beginning second history of computing. This paper introduces a new world model. 2 (section 19.6) is needed for simultaneously developing the scene toward parallel programming for manycore architectures and to structural programming for reconfigurable computing (RC), as well as heterogeneous systems including the cooperation of both paradigms. Figure 1: Lufthansa Reservation anno 1960 116 Currently the dominance of the basic computing paradigm is gradually wearing off (see section 19.4) with the growth of the area of Reconfigurable Computing (RC) applications -bringing profound changes to the practice of scientific computing, cyber physical systems (CPS) and ubiquitous embedded systems, as well as new promises of disruptive new horizons for affordable very high performance computing. Due to RC also the desk-top personal supercomputer is near. To obtain the payoff from RC we need a new understanding of computing and supercomputing, as well as of the use of accelerators (section 19.6.3). For bridging the translational gap, the software / configware chasm, we need to think outside the box3. Why Computers are Important Computers are very important for all of us. By many millions of people around the world computers are used everywhere. Typical orders of magnitude in the computer application landscape are: hundreds of applications, consisting of ten-thousands of programs, with millions of lines of code, having been developed by expenditure of thousands of man-years investment volumes up to billions of dollars3. We must maintain these important infrastructures. Wiki "Answers pages" nicely tell us, why computers running this legacy software are indispensable in the world 115 . The Computer is an electronic device used in almost every field even where it is most unexpected 117 . Now we cannot imagine a world without computers. These days' computers are the tools for not only engineers and scientists but also they are used by many millions of people around the world. The computer has become very important nowadays not only because it is accurate, fast and can accomplish many tasks easily. Otherwise to complete many tasks manually much more time is required (Fig. 1) . It can do very big calculations in just a fraction of a second. Moreover it can store huge amount of data in it. We also get information on many different aspects using internet on our computer. But there are more reasons, why computers are important. Many more different kinds of local or even worldwide infrastructures will be controlled by networks of computers 117 . Here computer crashes or software crashes may cause widely spread disasters by domino effects 89-95 . In his novel Hermann Maurer depicts the worldwide total chaos caused by a network crash in the year 2080 where millions of people die and the life of billions is threatened 90-91 . A thrilling novel? Yes, but also a textbook about possible solutions. 3
doi:10.1002/9781118342015.ch18 fatcat:shfb4oycu5hu5boizx6oltlgwa