J. A. Tierno. An Energy-Complexity model for VLSI computations. Ph.D. thesis, California Institute of Technology, 1995.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....As a result, the energy consumption of di erent operations can vary widely. Second, a metric that does not depend on the time taken by the operation (unlike power, for instance) allows us to compare di erent architectures that perform the same computation independent of timing considerations [32]. Tierno derived an elegant method by which we can estimate the energy per operation of the high level speci cation of an asynchronous circuit [32] We also showed how the energy per operation of the speci cation of a circuit can be related to the information theoretic entropy of its speci cation ....

.... (unlike power, for instance) allows us to compare di erent architectures that perform the same computation independent of timing considerations [32] Tierno derived an elegant method by which we can estimate the energy per operation of the high level speci cation of an asynchronous circuit [32]. We also showed how the energy per operation of the speci cation of a circuit can be related to the information theoretic entropy of its speci cation [33] We also derived lower bounds on the energy per operation by designing optimal coding techniques to specify the traces of the computation ....

Jose A. Tierno. An Energy Complexity Model for VLSI Computations. Ph.D. thesis, California Institute of Technology, 1995.

.... Besides the elimination of the clock skew and synchronization failure problems [34] asynchronous circuits also are more tolerant to variations in physical parameters, can be more easily synthesized using systematic and modular approaches [31] have a higher potential for low energy computation [42], and yield average case instead of worst case performance [24] The concept of asynchronous circuits has been around since the fifties [18] However, it has not gained popularity until recently because of the difficulties involved in removing hazards from early designs [44] Since then, several ....

J.A. Tierno. An Energy Complexity Model for VLSI Computations. Ph.D. thesis, California Institute of Technology, 1995.

....additive: we can calculate the index corresponding to an operation by adding the indices of all of its sub operations. 2.2 Synchronization We present next the part of the energy model that relates to the cost of synchronization between CSP statements. The cost of data operations is explained in [10]. The synchronization primitives of CSP are parallel composition ( k ) sequential composition ( guarded choice ( repetition, and bullet synchronization between communication actions ( ffl ) Some of these primitives have zero energy cost, such as parallel composition. Some of these ....

....be the length of the longest loop, counted as the number of selections made in that loop, and n be the number of statements in P. Then we can write k(HP; a 1: m ) K Theta m; therefore, C(HP; a 1: m ) is bounded above by Kdlog 2 ne, and the lim sup converges. Details of this proof can be found in [10]. The entropy of the statement sequences of length m, H(A 1 ; Am ) is defined in the usual way [9] 2 limsup n 1 an = lim n 1 sup i n (a i ) Definition 2 Let (A 1 ; Am ) be a sequence of random variables. The entropy of this sequence, H(A 1 ; Am ) is defined ....

[Article contains additional citation context not shown here]

Jos'e A. Tierno. An Energy Complexity Model for VLSI Computations. PhD thesis, California Institute of Technology, 1995.

No context found.

J. A. Tierno. An Energy-Complexity model for VLSI computations. Ph.D. thesis, California Institute of Technology, 1995.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC