A high-performance CMOS programmable logic core for system-on-chip applications

dc.contributor.authorHan, Yi, 1977-en_US
dc.date.accessioned2009-10-06T00:11:08Z
dc.date.available2009-10-06T00:11:08Z
dc.date.issued2005en_US
dc.descriptionThesis (Ph. D.)--University of Washington, 2005.en_US
dc.description.abstractIntegrated circuits (ICs) have experienced an astonishing revolution in the last two decades. As predicted by Moore's law, the integration density doubles approximately every eighteen months. Nowadays hundreds of millions of transistors can be integrated on a chip. To lower the cost, system-on-a-chip (SoC) has emerged as a solution and attracted attention. SoC technology is the packaging of all the necessary functionality, including digital, analog and radio frequency, on a single chip. It has been the hot spot of research for both industry and academic and is becoming the mainstream of IC design.As part of the SoCs, programmable logic cores (PLCs) have become more and more popular for its capability of post-fabrication changes. PLCs provide better flexibility, less cost and shorter time-to-market as well. On the other hand, the post fabrication flexibility sacrifices speed, resulting in a speed gap between fixed logic and programmable logic, and making programmable logic far from mainstream. Typically, a circuit implemented in programmable logic will run about 3 times slower than that implemented in fixed logic [8]. Improving the speed of programmable logic is one of the primary challenges.In this dissertation, a novel high-speed PLC architecture is proposed. By combining a high-performance dynamic logic style (output prediction logic or OPL), a product-term-based structure, mixed logic/routing structures, wired-OR structures and a unidirectional routing flow, this architecture achieves an average speedup of 4.1 times over the Xilinx Virtex-E FPGA [50], a Look-Up-Table (LUT) based architecture using static CMOS. It also can achieve up to a 1 GHz throughput, which is 2.4 times that of the Xilinx CoolRunner-II CPLD [49] and 3.2 times that of the Xilinx Virtex-E FPGA (Speed Grade-7) [50]. Measurement results for a prototype block fabricated in the TSMC 0.18mum/1.8V CMOS process are presented.This thesis work demonstrates that programmable logic can achieve comparable speeds to ASICs with design innovations, and therefore enables the adoption of programmable logic on an SoC.en_US
dc.format.extentx, 131 p.en_US
dc.identifier.otherb53683766en_US
dc.identifier.other61266954en_US
dc.identifier.otherThesis 54591en_US
dc.identifier.urihttp://hdl.handle.net/1773/5948
dc.language.isoen_USen_US
dc.rightsCopyright is held by the individual authors.en_US
dc.rights.urien_US
dc.subject.otherTheses--Electrical engineeringen_US
dc.titleA high-performance CMOS programmable logic core for system-on-chip applicationsen_US
dc.typeThesisen_US

Files

Original bundle

Now showing 1 - 1 of 1
Loading...
Thumbnail Image
Name:
3163380.pdf
Size:
10.65 MB
Format:
Adobe Portable Document Format