190 likes | 210 Views
FAULT-TOLERANT TECHNIQUES FOR NANOCOMPUTERS. GARIMELLA SUDHEER EE 585: Fault Tolerant Computing. NANOCOMPUTERS. A Nanocomputer is a computer whose physical dimensions are microscopic.
E N D
FAULT-TOLERANT TECHNIQUES FOR NANOCOMPUTERS GARIMELLA SUDHEEREE 585: Fault Tolerant Computing Fault Tolerant Computing
NANOCOMPUTERS • A Nanocomputer is a computer whose physical dimensions are microscopic. • The proposed Nanometer-Scale devices should eventually permit very large scale of integration of the order of 10^12 devices per chip. ( However this is still at academic level). • Current state-of-the-art microprocessors have more than 40 million transistors; by 2012 they could have five billion. Fault Tolerant Computing
Nanocomputers contd. • Nanoelectronic devices are more fragile than the conventional devices, and will be sensitive to external influences such as • Radiation related effects • High Temperature • Electromagnetic Interference • Parameter Fluctuations etc. • Hence, fault-tolerant architectures will certainly be necessary in order to produce reliable systems that are immune to manufacturing defects and to transient errors. Fault Tolerant Computing
Introduction • The nanometer-sized electronic devices are prone to errors at both Manufacturing and service level. • The present day manufacturing failure rate of CMOS is approximately 10^-7 to 10^-6. • Failure rate for highly experimental devices such as molecular transistors is currently just below unity. • Failure rate cannot be exactly predicted for nanoscale devices but it will be significantly poorer than the present day transistors. Fault Tolerant Computing
Fault Tolerant Techniques • The proposed fault tolerant techniques for nanocomputers are • R- Fold modular redundancy (RMR) • Cascaded triple modular redundancy (CTMR) • NAND Multiplexing • Reconfiguration. • The first two techniques are mainly used to correct transient errors. • Finally reconfiguration technique is suitable for tackling manufacturing defects. Fault Tolerant Computing
R-Fold Modular Redundancy • RMR is a generalization of TMR. • In Triple modular redundancy three units work in parallel and their outputs are compared with a majority gate. • Similarly in RMR R units work in parallel ( where R = 3,5,7,9…..). Fault Tolerant Computing
Cascaded Triple Modular Redundancy • The TMR process can be repeated by combining three of the TMR ‘units’ with another majority gate to form a ‘second order’ TMR unit with even higher reliability Fault Tolerant Computing
NAND Multiplexing • NAND multiplexing was proposed by Von Neumann 50 years ago. • This method is similar to RMR, but instead of majority gate to decide output, the output is carried out on a bundle of wires (Given as Nbundle ). • IT has two stages • Executive stage. • Restorative stage. Fault Tolerant Computing
NAND Multiplexing Contd. • The executive stage, performs the basic function of the processing unit in parallel. • The restorative stage, reduces the degradation caused by the executive stage and thus acts as a non-linear ‘amplifier’ of the output. • The signals to and from units are not carried in single lines but in bundles and the unit is replicated the appropriate number of times. Fault Tolerant Computing
NAND Multiplexing Contd. • If the inputs and processing units are perfectly reliable then the lines comprising each output should be identically stimulated (1) or unstimulated (0). • However, due to errors in the input data as well as errors occurring in the processing of the inputs from faulty devices, not all of the output lines in each output will be identically stimulated. • Thus, for multiplexed networks, the final outputs are considered to be 1 if more than (1 −a)Nbundlelines are stimulated and 0 if less than aNbundlelines are stimulated. ( 0<a<0.5, is predefined critical level). Fault Tolerant Computing
Reconfiguration • This technique is used for overcoming Manufacturing defects. • Defective devices are assembled in groups to form CLBs (Configurable logic blocks). • These CLBS are grouped together to form AFTBS (Atomic fault-tolerant blocks) • The AFTBs are then grouped together, NA at a time, to form a cluster which then performs some desired function. Fault Tolerant Computing
Results Fault Tolerant Computing
Emerging architectures for nanocomputers Fault Tolerant Computing
Conclusions • The implications of these results are that the future usefulness of various nanoelectronic devices may be seriously limited if they cannot be made in large quantities with a high degree of reliability. • It is theoretically possible to make very large functional circuits, even with one dead device in ten, but only if the dead devices can be located and the circuit reconfigured to avoid them. Fault Tolerant Computing
Current Nanocomputers • Spray on Nanocomputer • Teramac custom computer which uses Reconfiguration technique. Fault Tolerant Computing
Referances • Fault tolerant techniques for Nanocomputers by K Nikolic, A Sadek and M Forshaw. • Fault tolerant techniques for Nanocomputers: NAND Multiplexing By Jie Han and Pieter Jonker. • http://www.hpl.hp.com/personal/Bruce_Culbertson/TeramacBib.html Fault Tolerant Computing
QUESTIONS ?? Fault Tolerant Computing