RAMPRASAD, S., SHANBHAG, N. R., AND HAJJ, I. N. 1997. Analytical Estimation of Signal Transition Activity from Word-Level Statistics. IEEE Transactions on CAD of Integrated Circuits and Systems 16, 7 (Jul), 718--733.  Home/Search   Document Details and Download   Summary   Related Articles   Check

....e where: t t = 2E i Db i t e = E (2Db i i 1 i 1 ) # are respectively the temporal and the equivalent spatial transition activity. b represents the current value of the bus signal. In a wide spectrum of DSP applications the data distribution can be assumed to be Gaussian [2] with correlation factor r and variance s . In this case, the total signal activity T t and T e can be very accurately estimated by: T t K2 B 2 dtm log 2 (k s n ) b T others 0.5 log 2 (k s) b with parameters: k b g T t 1.90 0.84 2 T e 2.15 1.15 2.5 s n = s B t ....

S. Ramprasad, N. R. Shanbhag, and I. N. Hajj. Analytical estimation of signal transition activity from word level statistics. IEEE Trans. on CAD, pages 718 -- 733, July 1997.

....potentially excluding good designs [13] Existing work in data characterization, for the most part, has been geared towards illustrating and addressing special cases [9] 12] There are only a couple models that practically relate data characteristics to transition activity. Others, discussed in [17], 18] are limited by difficulty of use or constraining assumptions. One practical model is the Dual Bit Type (DBT) model [10] 11] which uses probabilistic parameters to characterize the data related activity. This model was developed to determine ranges for two types of activity that could be ....

S. Ramprasad, N. R. Shanbhag, and I. N. Hajj, "Analytical Estimation of Signal Transition Activity from Word-Level Statistics", IEEE Trans. Computer-Aided Design of Integrated Circuits and Systems, vol. 16, pp. 718-733, July 1997.

....good designs [10] 11] Existing work in data characterization, for the most part, has been geared toward illustrating and addressing special cases [12] 16] There are only a couple models that practically relate data characteristics to transition activity in general. Others, discussed in [17] [19] are limited by difficulty of use or constraining assumptions. One of the more practical models is the dual bit type (DBT) model [7] 14] which uses probabilistic parameters to characterize the data related activity. This model was developed to determine ranges for two types of activity ....

S. Ramprasad, N. R. Shanbhag, and I. N. Hajj, "Analytical estimation of signal transition activity from word-level statistics," IEEE Trans. Computer -Aided Design, vol. 16, pp. 718--733, July 1997.

....down the design space (i.e. without throwing away good design possibilities. Relationships between data characteristics and ff 0 1 in the datapath were first illustrated for special cases [3] 4] 5] Then, more general models for relating data characteristics and ff 0 1 began to appear [6] [7], 8] However, these models are targeted for high level power estimation accuracy (under certain assumptions) and do not analyze the causes of ff 0 1 . More general relationships need to be defined and modeled specifically to aid in finding low ff 0 1 high level designs. The relationships ....

S. Ramprasad, N. R. Shanbhag, and I. N. Hajj, "Analytical Estimation of Signal Transition Activity from Word-Level Statistics", IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, vol. 16, July 1997, pp. 718-733.

....can be taken in early design stages leading to cost reduction and shorter design cycles. The dominant factor of power dissipation in CMOS circuits is the dynamic power dissipation, which is directly proportional to the signal transition activity. Different terms such as transition activity [8,9], transition density [6] switching activity and transition probability [3,4] have been proposed to express the number of signal transitions. The objective is to establish an accurate model to predict the bit level switching activity from word level statistics, thus permitting the estimation of ....

....on signal magnitude and correlation factor. A data processing module is divided into parts characterized by different activities and the power consumption is estimated through the calculation of different capacitive coefficient vectors for each part of the module. The methodology proposed in [8,9] employs high level signal statistics, a statistical signal generation model and the signal encoding for the estimation of transition activity. The DBT model is used and an accurate relation for transition activity is derived when the bit level probability and correlation factor are known. Also, ....

[Article contains additional citation context not shown here]

Ramprasad, S., Shanbhag, N.R., and Hajj, I.N. "Analytical Estimation of Signal Transition Activity from Word-Level Statistics," IEEE Trans. on CAD, Vol.16, No.7, July 1997, pp.718-733.

....of bits in typical audio, video, and communications channel data is shown in Figure 2, from which, we see that the ]ess sig nificant bits (by ]ess significant bits, we mean bits close to and including the ]east significant bit) in a typical input sample have a transition activity close to 0. 5 [4, 6]. ttence, 0.55 atmlan 0.5 . 0.45 0.05 [igure 2. ansiion acivi versus bi position addresses formed by grouping bis in he less signcan bi positions re unormly distributed. From Figure 2, we so see ha he more sigca bis (by more signcan bis, we mean he bis close o and includin[ can ....

S. Ramprasad, N. R. Shanbhag, and I. N. Hajj, "Analytical Estimation of Signal Transition Activity from Word-Level Statistics," IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, vol. 16, no. 7, pp. 718 733, July 1997.

....of bits in typical audio, video, and communications channel data is shown in Figure 2, from which, we see that the less significant bits (by less significant bits, we mean bits close to and including the least significant bit) in a typical input sample have a transition activity close to 0. 5 [4, 6]. Hence, 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0 2 4 6 8 10 12 14 16 Bit atmlan audio3 audio4 audio5 audio6 audio7 video3 Figure 2. Transition activity versus bit position addresses formed by grouping bits in the less significant bit positions are uniformly distributed. From Figure ....

S. Ramprasad, N. R. Shanbhag, and I. N. Hajj, "Analytical Estimation of Signal Transition Activity from Word-Level Statistics," IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems, vol. 16, no. 7, pp. 718--733, July 1997.

....from word level signal properties is described in section 2.2. In section 2.3, we compute transition activity for various filter structures and in section 2.4 we present simulation results for audio, video, and communication system signals and filters. The results of this chapter have appeared in [28, 29, 30]. 2.1 Preliminaries In this section, we will present definitions and review existing results that will be employed in later sections. First, we will define the word level quantities such as the mean ( variance (oe 2 ) and temporal correlation (ae) Next, we consider bit level quantities ....

....of bits in typical audio, video, and communications channel data is shown in Figure 3.23, from which, we see that the less significant bits (by less significant bits, we mean bits close to and including the least significant bit) in a typical input sample have a transition activity close to 0. 5 [14, 30]. Hence, addresses formed by grouping bits in the less significant bit positions are uniformly 0 0.05 0.1 0.15 0.2 0.25 0.3 0.35 0.4 0.45 0.5 0.55 0 2 4 6 8 10 12 14 16 Bit atmlan audio3 audio4 audio5 audio6 audio7 video3 Figure 3.23 Transition activity versus bit position distributed. From ....

[Article contains additional citation context not shown here]

S. Ramprasad, N. R. Shanbhag, and I. N. Hajj, "Analytical Estimation of signal transition activity from word-level statistics," IEEE Transactions on Computer-Aided Design of ICs, pp. 718--733, vol. 16, no. 7, July 1997.

No context found.

RAMPRASAD, S., SHANBHAG, N. R., AND HAJJ, I. N. 1997. Analytical Estimation of Signal Transition Activity from Word-Level Statistics. IEEE Transactions on CAD of Integrated Circuits and Systems 16, 7 (Jul), 718--733.

No context found.

S. Ramprasad, N. R. Shanbhag, and I. N. Hajj, "Analytical Estimation of Signal Transition Activity from Word-Level Statistics," IEEE Trans. on CAD, vol. 16, no. 7, pp. 718-733, Jul 1997.

No context found.

RAMPRASAD,S.,SHANBHAG,N.R.,AND HAJJ, I. N. 1997. Analytical Estimation of Signal Transition Activity from Word-Level Statistics. IEEE Transactions on CAD of Integrated Circuits and Systems 16, 7 (Jul), 718--733.

No context found.

S. Ramprasad, N. Shanbhag, and I. Hajj. Analytical Estimation of Signal Transition Activity from Word-Level Statistics. IEEE Transactions on CAD, 16(7):718--733, July 1997.

No context found.

S. Ramprasad, N. R. Shanbhag, and I. N. Hajj, "Analytical Estimation of Signal Transition Activity from Word-Level Statistics," IEEE Transactions on CAD of Integrated Circuits and Systems, vol. 16, no. 7, pp. 718--733, Jul 1997.

No context found.

S. Ramprasad, N. R. Shanbhag, and I. N. Hajj, "Analytical Estimation of Signal Transition Activity from Word-Level Statistics," IEEE Trans. on CAD, vol. 16, no. 7, pp. 718-733, Jul 1997.

No context found.

S. Ramprasad, N. Shanbhag and I. Hajj, "Analytical Estimation of Signal Transition Activity from Word-Level Statistics," IEEE Transactions on Computer Aided Design of Integrated Circuits and Systems, Vol 16, no. 7, pp. 718-733, July 1997.

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