flickr - AMagill
flickr - AMagill
Comparisons
We focus on other 2 kinds of systems comparing to the Reconfigurable Computing System (RCS). The first one contains high-end processor. The main advantage it has is that it has great information processing power such as the Intel i7-960 which runs at 12.8 GHz for the whole unit but the power required is about 130 Watts and it is out of reach for normal customer because it is over $560 US dollars (2010). There is another type of system to improve the performance, namely Application-Specific Integrated Circuits (ASICs). It just has the right amount of function and silicon area to do the targeted task. It is fast and effective due to the less sufferance from the instruction fetching, encode and decode data process. As a source1 pointed out, it still has 2 main problems. The first one is the daunting overhead, and the other is the unacceptable time during the designing period. Finally if we look up the RCS, the core of the system can execute code effectively due to the system that has been evolved to its optimal state and like the ASICs, it also contain the right mix of content in the system and yet the system does not need to be completely redesigned. Only the parts of time-critical and has a high degree of parallelism are needed to remapped. This means that it is as effective as the ASICs and yet it has much more flexibilities than ASICs due to its abilities to evolve into other target-specific circuits. Another source2 has pointed out that performing multiplication on the XC2V6000 FPGA with 66 MHz is 540 times faster than a dual-core CPU running at 2.6 GHz. Thus, it is possible to predict that RCS may overrun the systems with high-end processor.
Two particular cases of RCS are Reconfigurable Field Programmable Analogue Array (FPAA) and Field Programmable Gate Array (FPGA). FPAA is used for building analogue circuit and FPGA is for the digital circuit design. If we compare them with the progress of circuit has developed, it seems FPGA is much further developed. The reason that it is harder to build an analogue circuit on the board is due to the difficulties of evaluating the parameters3 such as frequency, voltage gain and transient response. While in the digital domain, we can just choose the parameters4 as bits to alter the properties of gates. Hence, our main focus is on the system built on the FPGA board in the following section.
References
1 Todman,T.J., Constantinides, G.A., Wilton, S.J.E., Mencer, O., Luk, W., Cheung., P.Y.K.: Reconfigurable computing: architectures and design methods. IEE PROCEEDINGS. COMPUTERS AND DIGITAL TECHNIQUES 152 (2) pp. 193-207 (2005)
2 Telle, N., Cheung, C.C., and Luk, W.: Customising hardware designs for elliptic curve cryptography’, Lect. Notes Comput. Sci., 2004, 3133 (2004)
3 Greenwood, G.W., Tyrrell, A.M.: Introduction To Evolvable Hardware: A Practical Guide for Designing Self-Adaptive Systems. pp. 67 (2006)
4 Wang, J.: Design and implementation of a virtual reconfigurable architecture for different applications of intrinsic evolvable hardware, IET COMPUTERS AND DIGITAL TECHNIQUES 2 (5) pp. 386-400 (2008)