Web Designer and Developer Gavin Stewart

Computer Architectures

The development of VSLI CMOS systems will continue with new silicon based nanotechnologies, such as silicon-on-insulator, for at least another decade and should allow for the development of systems that will operate at speeds of up to 15GHz. There are however issues with CMOS systems that will mean they cannot be developed indefinitely. These issues are set out in the SIA roadmap, which identifies several major technical difficulties in the development of CMOS systems. These are power management, new architectures to overcome bottlenecks at interconnects, ultimate short channel limitations requiring more complex gate structures such as SOI or dual-gate transistors, the spiralling costs of lithography and fabrication (Beckett and Jennings 2002). Almost all VSLI current computer architectures are based on the stored-program (Von Neumann) architecture that separates out the processing and the computer memory then communicates between them through busses (Figure 5). It will be impossible for the stored program architecture to continue when communication between even adjacent switches is difficult (Beckett and Jennings 2002). If nanotechnology is to continue to provide improved performance within computers then nano-computer architectures must supersede conventional architectures, while performing the same high level functions as their predecessors as well as sharing many of their characteristics.

Whereas current computer architectures rely on computation to be error free nano-computer architectures will have high rates of failure and varying performance levels (Beckett 2003). This is because “physical problems such as leakage, threshold voltage, tunnelling, electro-migration, high interconnect resistance, crosstalk and the need for robust and flexible error management become significant as device features shrink” (Beckett and Jennings 2002). These problems can affect the way devices are connected together and will cause future computer architectures to move towards locally-connected, reconfigurable hardware meshes that merge processing and memory (Beckett and Jennings 2002). The biggest challenge in nano-electronics has been described as the development of logic designs and computer architectures necessary to link small, sensitive devices together to perform useful calculations efficiently (Beckett and Jennings 2002).

Figure 5 – A representation of the Von Neumann architecture (Wikipedia 2005).

There are several possibilities for new architectures including, Synthetic Neural Systems, Quantum Dot Arrays and Locally Connected Machines. One that seems to show promise at the moment is Quantum Dot Arrays. Quantum Dot Arrays or “Quantum Dot Automata is a novel computing mechanism that can represent binary information based on spatial distribution of electron charge configuration in chemical molecules” (Racichandran et al 2004). Figure 6 shows a high level Quantum Dot Array of four dots that can be used to represent the binary value’s 0 and 1 (Neimier et al 2000). The array contains four quantum dots and two electrons. The electrons can be moved between the quantum dots using electron tunnelling. At the different positions shown in Figure 6 they represent the binary values 1 and 0 encoded in the charge configuration within the quantum dot cells (Neimier et al 2000). These QCA cells can then be combined together to make simple logic devices. These devices include QCA wires, the majority gate, and other more complex devices (Neimier et al 2000). Some simple QCA logic devices are represented in Figure 7. These simple devices can be used to create more complex logic device to use as AND gates, OR gates or inverters (Racichandran et al 2004). These logic gates can then be used to create a processor. A simple QCA processor architecture has already been proposed – called Simple 12. A high level diagram of Simple 12 can be seen in Figure 8. Simulation tests on Simple 12 have shown that it is a “theoretically correct QCA design” (Neimier et al 2000).

Before the QCA Simple 12 was designed a CMOS version was also designed. The comparisons between these showed that “the QCA offers at least an order of magnitude area density increase over the equivalent CMOS design.” (Neimier et al 2000). Additionally, scientists are currently working to reduce the size of QCA cells so that “potential density gains are three orders of magnitude.” (Neimier et al 2000).

Although, the CMOS systems and the QCA system are built very differently they work in a similar way at a high level and they both use binary representations of data. The diagram of the Von Neumann architecture in Figure 5 when compared to the QCA diagram in Figure 8 shows that the two architectures have similar features, such as logic units and clocks, however the QCA architecture does not contain any long busses over which data has to be transmitted.

Figure 6 - QCA cell polarisation and representations of binary 1 and 0 (Neimier et al 2000).

Figure 7 - (a) Majority Gate, (b) 90 degree wire, (c) Wire crossing (Racichandran et al 2004).

Figure 8 – A diagram of Simple 12: a possible QCA architecture (Neimier et al 2000).

Risks with Nanotechnologies

Much has been made of the perceived risks posed by the development of nanotechnologies; however the actual risks are minimal. One of the most famous examples is the “grey goo” scenario proposed by Michael Crichton where nano-robots uncontrollably self replicate destroying everything in their path. These stories are however “more in the tradition of King Kong than the realms of scientific plausibility” (Donaldson and Stone 2004). Other worries include nano materials damaging vital organs such as the digestive system or the lungs through inhaling particles. As Donaldson and Stone (2004) explain there is a potential risk to the body from nanoparticles, however, nanotechnologies are successfully in use today without any health risks to the people working with them and “not surprisingly, public fears are directly correlated with the amount of knowledge that people have about nanotech: the less knowledge, the more fear” (Lawrence 2005) as is shown in Figure 9.

Figure 9 – Chart showing peoples perception of nanotechnology risks (Lawrence 2005).

Conclusion

Nanotechnologies are contemporary technologies that are already in commercial use as well as in research and development. Nanotechnologies are already affecting the computing industry and will continue to do so as they are further developed in the future. Nanotechnologies will continue to improve computing as we know it for at least another decade, but the development of VSLI CMOS systems will reach its limit and a new type of nano-computer will be needed to replace it. These systems will use very different computer architectures from today, possible not even electronic, but will to work in the same way as they do currently at a high level and support inherently linear software development and legacy systems. An architecture that shows particular promise is QCA, which could be a very viable alternative to CMOS (Neimier et al 2000). Where the speed of CMOS systems will be limited to around 15GHz these new architectures could provide “operating frequencies in excess of 100GHz” (Beckett and Jennings 2002). Although there are minimal risks with nanotechnologies, the possible advantages that can be gained far outweigh any risks.

References

• Beckett, Paul. 2003. Exploiting multiple functionality for nano-scale reconfigurable systems. Great Lakes Symposium on VLSI, Washington, 28-29 April. [online] New York: ACM Press, pp.50-55. Available from: http://portal.acm.org/ft_gateway.cfm?id=764822&type=pdf&coll=ACM&dl=ACM&CFID=4145846&CFTOKEN=70281767 [Accessed 10 November 2006]
• Beckett,P and Jennings A. 2002. Towards nanocomputer architecture. ACM International Conference Proceeding Series, Melbourne. [online] Darlinghurst: Australian Computer Society, Inc, pp.141-150. Available from: http://portal.acm.org/ft_gateway.cfm?id=563950&type=pdf&coll=ACM&dl=ACM&CFID=2270945&CFTOKEN=37058276 [Accessed 10 November 2006]
• Donaldson,K and Stone, V. 2004. Nanoscience fact versus fiction. Communications of the ACM. 47(11): pp.113-115. [online]. Available from: http://portal.acm.org/ft_gateway.cfm?id=1029518&type=pdf&coll=ACM&dl=ACM&CFID=2270945&CFTOKEN=37058276 [Accessed 10 November 2006]
• IBM.Photograph. [online image]. Ibm.com. Available from: http://www-03.ibm.com/ibm/history/exhibits/vintage/vintage_4506VV1003.html [Accessed 12 December 2006]
• Instituteof Nanotechnology. 2006. [online]. Nano.org.uk. Available from: http://www.nano.org.uk/whatis.htm [Accessed 30 October 2006]
• Intel.2006. Chart. [online image]. Intel.com. Available from: http://www.intel.com/technology/mooreslaw/index.htm [Accessed 12 December 2006]
• Lawrence,S. 2005. Nanotech Grows Up. Technology Review. 108(6): pp.31-31. [online]. Available from: http://search.ebscohost.com/login.aspx?direct=true&db=bsh&AN=17213257&site=ehost-live [Accessed 25 November 2006]
• Lundstorm,Mark. 2002. Is nanoelectronics the future of microelectronics? International Symposium on Low Power Electronics and Design, Monterey, 12-14 August. [online] New York: ACM Press, pp.172-177. Available from: http://portal.acm.org/ft_gateway.cfm?id=566410&type=pdf&coll=ACM&dl=ACM&CFID=2270945&CFTOKEN=37058276 [Accessed 10 November 2006]
• Neimier,M T et al. 2000. A design of and design tools for a novel quantum dot based microprocessor. Annual ACM IEEE Design Automation Conference, Los Angeles, 05-07 June. [online] New York: ACM Press, pp.227-232. Available from: http://portal.acm.org/ft_gateway.cfm?id=337398&type=pdf&coll=ACM&dl=ACM&CFID=4145846&CFTOKEN=70281767 [Accessed 10 November 2006]
• OxfordInstruments. 2006. Chart. [online image]. Oxinst.com. Available from: http://www.oxinst.com/wps/wcm/resources/file/ebf9b00f461238a/Nanotech_brochure_2006.pdf [Accessed 25 November 2006]
• Racichandran,R et al. 2004. Automatic cell placement for quantum-dot cellular automata. Great Lakes Symposium on VLSI, Boston, 26-28 April. [online] New York: ACM Press, pp.332-337. Available from: http://portal.acm.org/ft_gateway.cfm?id=989033&type=pdf&coll=ACM&dl=ACM&CFID=4145846&CFTOKEN=70281767 [Accessed 15 November 2006]
• RoyalSociety. 2006. [online]. RoyalSoc.ac.uk. Available from: http://www.royalsoc.ac.uk/landing.asp?id=1210 [Accessed 30 October 2006]
• TheNanoTechnology Group Inc. 2006. [online] TheNanotechnologyGroup.org. Available from: http://www.thenanotechnologygroup.org/index.cfm?content=118 [Accessed 10 November 2006]
• Wikipedia.2005. Illustration. [online image]. Wikipedia.org. Available from: http://en.wikipedia.org/wiki/Von_Neumann_architecture [Accessed 19 November 2006]
• Wikipedia.2006. [online]. Wikipedia.org. Available from: http://en.wikipedia.org/wiki/Nanotechnology [Accessed 30 October 2006]
• Winters,J. 2005. Are we there yet? Mechanical Engineering. 127(4): pp.4. [online]. Available from: http://search.ebscohost.com/login.aspx?direct=true&db=bsh&AN=16712955&site=ehost-live [Accessed 15 November 2006]

Download the whole report on the future of nanotechnologies in computing in PDF form