Asynchronous Microprocessor the Clockless Future
نویسنده
چکیده
distributing the clock signal on the chip in a manner such that it reaches all sinks at exactly the same point in time, a huge effort must be made. For example, the H-clock tree or B-Tree can achieve a low clock skew at the cost of many additional wires, which increases the chip size, cost and energy dissipation due to the capacitances that are created by thousands of parallel wires in the clock tree. The speeding up of a synchronous processor is possible by optimising the critical path. Either its elements have to be optimised or it has to be split up so that it takes more than one clock cycle to compute the final result. The latter way introduces a new overhead. But, in asynchronous processors, there are several starting points: Each optimised element can speed up the processor, so there is no need to make effort to optimise only rarely used parts. Instead, only frequently used parts are optimised to increase the average speed of the entire processor. If changes in combinatory logic elements (like increase in temperature and lowered supply voltage) slow down the operational performance of the processor, the clock of the synchronous chip has to be regulated actively to keep the system in a valid state. But, asynchronous systems can implicitly adapt to these variations, which means that they always run at the maximum speed possible. Another implication of the automatic performance adaptation is that an asynchronous system always World’s first flexible 8-bit asynchronous microprocessor developed by Seiko Epson using low-temperature polysilicon thin-film transistors on a plastic substrate. With energy consumption reduced by 70 per cent compared to the synchronous microprocessors now in everyday use, Epson is now researching potential applications for its invention (courtesy: http://global.epson.com)
منابع مشابه
25 Years Ago: the First Asynchronous Microprocessor
Twenty-five years ago, in December 1988, my research group at Caltech submitted the world’s first asynchronous (“clockless”) microprocessor design for fabrication to MOSIS. We received the chips in early 1989; testing started in February 1989. The chips were found fully functional on first silicon. The results were presented at the Decennial Caltech VLSI Conference in March of the same year. Th...
متن کاملPower-Efficient Design of a Clockless NoC Router with a New Integrated Flow
The downscaling of silicon technology and the capacity to build entire systems on a chip have made intrachip communication a relevant research topic. Besides, technology challenges point to the fast adoption of non-synchronous networks on chip (NoCs), using either globally asynchronous, locally synchronous (GALS) or even clockless approaches. However, clockless circuit design with current autom...
متن کاملYield Modeling and Analysis of a Clockless Asynchronous Wave Pipeline with Pulse Faults
This paper proposes a new fault model and its modeling and analysis methods in a clockless asynchronous wave pipeline for extensive yield evaluation and assurance. It is highly desirable to have an adequate and specific pulse fault rate model for establishing a sound theoretical foundation for clockless wave pipeline design for reliability. The pulse fault model is thoroughly identified as the ...
متن کاملThe MANGO ClocklessNetwork-on-Chip:*0.5exConcepts and Implementation
Shared, segmented, on-chip interconnection networks, known as networks-on-chip (NoC), may become the preferred way of interconnecting IP cores in future giga-scale system-on-chip (SoC) designs. A NoC can provide the required communication bandwidth while accommodating the effects of scaling microchip technologies. Equally important, a NoC facilitates a truly modular and scalable design flow. In...
متن کامل