T. L. Chou, K. Roy and S. Prasad, "Estimation of Circuit Activity Considering Signal Correlation and Simultaneous Switching," ICCAD-94: IEEE/ACM International Conference on Computer Aided Design, pp. 300-303, San Jose, CA, November 1994.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....nd MA models re special cses of ARMA models. An M tn order uto regressive (AR(M) signal model is identical to n ARMA(O,M) model. Also, n N tn order moving verge (MA(N) signal model is the sme s n ARMA(N,O) model. In proving Theorem i in section III, we will lso employ the following result from [14], Lemma 1: E[bi(n)bi(n 1) Pi 2 III. WORD LEVEL SIGNAL TRANSITION ACTIVITY In this section, we will present techniques for estimating word level transition activity, T, of a signal, x(n) from its word level statistics. We will first present a theorem relating bit level quantities, namely, ....

Tan-Li Chou, K. Roy, and S. Prasad, "Estimation of Circuit Activity Considering Signal Correlations and Simultaneous Switching," International Conference on Computer-Aided Design, pp. 300-303, November 1994.

....has been undertaken in the area of power consumption in CMOS (complimentary metal oxide semiconductor) devices. This work has focused primarily on circuit activity and on power consumption in circuits designed using VLSI basic cell techniques. Representative examples of this work are found in [3] [4], and [6] Besides [7] the closest research to power prediction in configurable devices is found in [6] where a power prediction tool was developed to predict power consumption in circuits that have been designed using VLSI techniques. This is similar to predicting power consumption in FPGAs, ....

T.L. Chou, K. Roy and S. Prasad, "Estimation of Circuit Activity Considering Signal Correlat ions and Simultaneous Switching," Proc. IEEE International Conference on Computer-Aided Design, pp. 300-303, Nov. 1994.

....performance (i.e. speed) is a main design consideration; however, power consumption is of growing concern as the logic density and speed of ICs increase. Some research has been undertaken in the area of power consumption in CMOS (complimentary metal oxide semiconductor) devices, e.g. see [4, 5]. However, most of this past work assumes design and implementation based on the use of standard (basic cell) VLSI techniques, which is typically not a valid assumption for application circuits designed for implementation on an FPGA. 2 Overview of the Tool A probabilistic power prediction tool ....

T. L. Chou, K. Roy, and S. Prasad, "Estimation of Circuit Activity Considering Signal Correlations and Simultaneous Switching," Proc. IEEE Int'l Conf. Comput. Aided Design, pp. 300-303, Nov. 1994.

....circuit problem , is by no means trivial. Indeed, a whole set of solutions have been proposed, ranging from approaches which build the global OBDDs [2] and therefore capture all internal dependencies, to efficient techniques which partially account for dependencies in an incremental manner [3] [5]. Recently, some authors have pointed out the importance of correlations not only inside the target circuit, but also at the circuit inputs. If one ignores these correlations, the power estimation results can be significantly impaired [6] 7] We will refer to this issue as the input problem and ....

T. L. Chou, K. Roy, and S. Prasad, `Estimation of Circuit Activity Considering Signal Correlations and Simultaneous Switching', in Proc. IEEE/ACM Intl. Conference on Computer Aided Design, pp. 300-303, Nov. 1994.

....uncertainty of a random variable with discrete probability distribution. The entropy of a binary random variable is defined as [3,7] Hpppp = log log) 0011 , 5) where log denotes the base 2 logarithm function. By definition the entropy function takes only positive values within the interval [0,1]. The minimum value 0 is obtained for p 0 or p 1 equal to 1, complete certainty, while the maximum value of 1 is obtained for p 0 = p 1 = 0.5, maximum uncertainty. The conditional entropy between two successive time steps, H c , expressing the uncertainty of b(n) i given that b(n 1) j is [3,7] ....

....( 13) The transformation (p 10 0. 5) 2 u leads to: 842 052 432 uuuu H c = 14) In general (14) has four complex roots which lead to eight complex roots of (13) However since 1 H c 0 there is always a real solution of (14) which leads to a value of p 10 within the interval [0,1]. Actually the proposed approximation is valid for H c 0.134 where ( pp p p p 10 10 10 1 10 10 1 . This is equivalent to 0.0185 p 10 0.9815 which is true for all real life signals, since the corresponding correlation factor is greater than 0.995 or less than 0.005. 3. BASIC ....

Chou, T.-L., Roy, K., Prasad, S. "Estimation of Circuit Activity Considering Signal Correlations and Simultaneous Switching," Proc. of ICCAD 94, pp.300-303.

.... BDDs can be used [25] to compute the Boolean difference probabilities, which are required in order to use the propagation algorithm (7) Recently, special 5 8 ized BDD based techniques have been proposed to facilitate this [39] Improvements on this basic technique have also been proposed in [31, 51], providing more accurate gate models and improved correlation handling. 3.2.2.4. A symbolic technique The technique proposed in [28] attempts to handle both spatial and temporal correlations by using a BDD to represent the successive Boolean functions at every node in terms of the primary ....

T-L. Chou, K. Roy, and S. Prasad, "Estimation of circuit activity considering signal correlations and simultaneous switching," IEEE/ACM International Conference on Computer-Aided Design, pp. 300--303, November 1994.

....circuit problem , is by no means trivial. Indeed, a whole set of solutions have been proposed, ranging from approaches which build the global OBDDs [2] 3] and therefore capture all internal dependencies, to efficient techniques which partially account for dependencies in an incremental manner [4] [7]. Recently, some authors have pointed out the importance of correlations not only inside the target circuit, but also at the circuit inputs. If one ignores these correlations, the power estimation results can be impaired [8] 9] We will refer to this as the input problem which is important not ....

T. L. Chou, K. Roy, and S. Prasad, `Estimation of Circuit Activity Considering Signal Correlations and Simultaneous Switching', in Proc. IEEE/ACM Intl. Conference on Computer Aided Design, pp. 300-303, Nov. 1994.

....The problem of determining when and how often transitions occur at a node in a general digital circuit is difficult because it is highly input dependent. As a result, researchers have resorted to stochastic approaches, where the probability of transition of all nodes in the circuit is estimated [3] [4] 6] 11] 13] 15] 17] This paper presents a hierarchical approach for power estimation of digital circuits at the transistor level. This is achieved by decomposing the digital circuit into sub circuits, with each sub circuit performing a particular operation. For example, a typical ....

Chou, T.-L., Roy, K., and Prasad, S. Estimation of circuit activity considering signal correlations and simultaneous switching. In Proc. ICCAD (1994), pp. 300--303.

....work done in this area involved performing exhaustive simulations for randomly generated input sequences [2] 9] Although these methods gave accurate results, lengthy computation times rendered them impractical for large circuits. As a result researchers have resorted to stochastic approaches [3], 6] 8] 10] 12] 14] 15] where the probability of transition of nodes in the circuit is estimated. This is because in a typical CMOS digital circuit with negligible leakage current, power is dissipated only when there is a transition from zero to one (or one to zero) at the node of a ....

T.-L. Chou, K. Roy, and S. Prasad. Estimation of circuit activity considering signal correlations and simultaneous switching. In Proc. ICCAD, pages 300--303, 1994.

.... and propagated through the netlist yielding the switching activity for each fanout node and thus the capacitance switching portion of power dissipation [2, 3] Signal correlations due to correlated inputs and the circuit structure can be taken into account in order to achieve more exactness [4, 5]. Explicit logic simulation techniques, which are in general more time intensive, require application typical binary patterns to be applied to the inputs. Using a suitable characterisation of standard cells and a simulation model for slopes and (partial) glitches, short circuit currents can be ....

T.-L. Chou, K. Roy, S. Prasad. Estimation of Circuit Activity Considering Signal Correlations and Simultaneous Switching. ICCAD'94, 1994, pp. 300-303.

....the algorithm does not consider simultaneous switching of multiple primary inputs. Hence, activity of a node is overestimated. In [5] Ghosh et al. presented a method that deals with simultaneous switching of multiple primary inputs but neglected temporal correlation between primary inputs. In [3], Chou et al. and in [12] Tsui et al. presented exact methods to calculate signal probability and activity. The above methods can be used to accurately estimate the power dissipation of a circuit provided that the exact probability and activity of each primary input is known. However, accurate ....

.... P m i;k (x k (t Gamma T ) P l i;k (x k (t Gamma T ) p X j=1 ff j ( n Y k=1 P m j;k (x k (t) P l j;k (x k (t) Delta 7 = X 1i;jp ff i ff j ( n Y k=1 P m i;k (x k (t Gamma T ) P l i;k (x k (t Gamma T ) P m j;k (x k (t) P l j;k (x k (t) 19) In [3] it has been shown that if P (x k (t Gamma T ) P (x k (t) is replaced with P (x k (t Gamma T )x k (t) P (x k (t Gamma T ) P (x k (t) with P (x k (t Gamma T )x k (t) P (x k (t Gamma T ) P (x k (t) with P (x k (t Gamma T )x k (t) and P (x k (t Gamma T ) P (x k (t) with P (x ....

Chou, T. L., Roy, K., and Prasad, S. Estimation of circuit activity considering signal correlations and simultaneous switching. IEEE Intl. Conf. on Computer-Aided-Designed (1994). pp.300-303.

....1 Introduction The increasing use of portable computing and communication systems makes power dissipation a critical parameter to be minimized in circuit design. Therefore, power estimation tools are badly needed. In order to accurately estimate power, traditional power estimation techniques [3, 4, 5, 6] require exact signal properties of primary inputs. However, accurate signal property values for primary inputs may not often be available. Since power dissipation strongly depends on the input signal properties, uncertainties in specifications of input signal properties make the estimation ....

T.-L. Chou, K. Roy, and S. Prasad, "Estimation of Circuit Activity Considering Signal Correlations and Simultaneous Switching," IEEE Intl. Conf. on Computer-Aided-Designed, pp.300-303, 1994.

....consists of dynamic and static components. In circuits with a high supply voltage, a relatively high transistor threshold voltage can be used, making subthreshold current negligible. That is the common assumption for the existing gate level techniques for average power estimation [1] 3] 4] [5] [6] 7] 14] 15] 16] 17] 18] 22] 23] However, low power applications have been driving the supply voltage to become lower and lower, which requires the device threshold to be reduced so as to satisfy performance requirements. Unfortunately, due to the exponential relationship between ....

T.L. Chou, K. Roy, and S. Prasad, "Estimation of Circuit Activity Considering Signal Correlations and Simultaneous Switching, " IEEE Intl. Conf. on Computer-Aided-Designed, pp.300303, 1994.

....dissipation a critical parameter to be minimized in circuit design [2] 13] Therefore, power estimation tools are of critical importance. Methods for estimation of power dissipation in combinational circuits include probabilistic and statistical techniques [11] In the probabilistic approach [6], 7] 8] 12] the signals are usually modeled as stochastic processes where each signal has properties such as signal probability (probability of a signal being logic ONE) and activity (probability of signal switching) A single symbolic simulation run determines the internal node signal ....

....symbolic simulation run determines the internal node signal probabilities and activities by propagating the input property information through the circuit. In order to accurately estimate average power dissipation, spatio temporal correlations among signals have to be taken into consideration [6], 9] 10] Statistical methods are basically Monte Carlo based techniques [1] 15] They simulate the circuits with a limited number of input vectors which are randomly generated conforming to the given probabilities and activities of the input signals. The number of simulation runs are ....

T.-L. Chou, K. Roy, and S. Prasad, "Estimation of Circuit Activity Considering Signal Correlations and Simultaneous Switching, " IEEE Intl. Conf. on Computer-Aided-Designed, pp.300303, 1994.

....circuit design [1] In order to determine the bottleneck for power and to synthesize circuits satisfying the power requirements, accurate estimation of power dissipation is needed. Recently, several probabilistic and statistical techniques have been proposed to estimate power dissipation [2] [3], 4] 5] 6] Some of them can be used to accurately estimate the power dissipation of a circuit provided the exact probability (probability of a signal being logic ONE) and activity (probability of signal switching) of primary inputs are known. However, accurate signal probability or activity ....

.... U (y(t) and U (y(t Gamma T ) represent the sums of probability products of y(t) and y(t Gamma T ) Hence, we have, U (y(t Gamma T ) U (y(t) X 1i;jp ff i ff j ( n Y k=1 P m i;k (x k (t Gamma T ) P l i;k (x k (t Gamma T ) P m j;k (x k (t) P l j;k (x k (t) 15) In [3] it has been shown that if P (x k (t Gamma T ) P (x k (t) is replaced with P (x k (t Gamma T )x k (t) P (x k (t Gamma T ) P (x k (t) with P (x k (t Gamma T ) x k (t) P (x k (t Gamma T ) P (x k (t) with P ( x k (t Gamma T )x k (t) and P (x k (t Gamma T ) P (x k (t) with ....

[Article contains additional citation context not shown here]

T.-L. Chou, K. Roy, and S. Prasad, "Estimation of Circuit Activity Considering Signal Correlations and Simultaneous Switching, " IEEE Intl. Conf. on Computer-Aided-Designed, pp.300303, 1994.

....the size of symbolic probability expression and activity expression grow exponentially with respect to the number of independent inputs. Consequently, we resort to circuit partitioning to trade off accuracy versus computation time. The circuit partitioning technique is similar to the one given in [4]. 4.2 Statistical Technique to Estimate Power Sensitivities The basic idea of the statistical technique is to obtain sensitivities as a by product of statistical power estimation processes. A more detailed derivation can be found in [3] The expected value of average power in a CMOS circuits can ....

T.-L. Chou, K. Roy, and S. Prasad, "Estimation of Circuit Activity Considering Signal Correlations and Simultaneous Switching," IEEE Intl. Conf. on Computer-Aided-Designed, pp. 300-303, 1994.

No context found.

T. L. Chou, K. Roy and S. Prasad, "Estimation of Circuit Activity Considering Signal Correlation and Simultaneous Switching," ICCAD-94: IEEE/ACM International Conference on Computer Aided Design, pp. 300-303, San Jose, CA, November 1994.

No context found.

T. L. Chou, K. Roy, S. Prasad, "Estimation circuit activity considering signal correlations and simultaneous switching," in Proc. IEEE ICCAD, pp.300-303, 1994.

No context found.

T-L. Chou , K. Roy and S. Prasad. " Estimation of circuit activity considering signal correlation and simultaneous switching. " In Proceedings of the IEEE International Conference on Computer Aided Design, pages 300-303, November 1994.

No context found.

Tan-Li Chou, Kaushik Roy, and Sharat Prasad. Estimation of circuit activity considering signal correlations and simultaneous switching. In IEEE/ACM International Conference on Computer-Aided Design, pages 300--303, 1994.

No context found.

T-L. Chou , K. Roy and S. Prasad. " Estimation of circuit activity considering signal correlation and simultaneous switching. " In Proceedings of the IEEE International Conference on Computer Aided Design, pages 300-303, November 1994.

No context found.

T. Chou, K. Roy, and S. Prasad, "Estimation of circuit activity considering signal correlations and simultaneous switching," in Proc. Int. Conf. Computer Aided Design, 1994, pp. 300--309.

No context found.

T.-L. Chou, K. Roy, and S. Prasad, "Estimation of circuit activity considering signal correlations and simultaneous switching", in Proc. ICCAD, pp. 300--303, 1994.

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