Information Technology Reference
In-Depth Information
The sub-threshold current is given by equation (1):
⁄
.
⁄
(1)
Here k and n are technology parameters, V
t
is the temperature voltage, V
th
is the
threshold voltage, V
gs
is the gate-source voltage and V
ds
is the drain-source voltage.
The equation shows that the leakage current is dependent on width. After estimating
the leakage current, we can evaluate the static power using equation (2)
.
(2)
The early estimation is useful for power management i.e., the power performance
can be evaluated prior to the actual design of architecture. Moreover, end users can
optimize their software as per the requirements. Thus, it is beneficial for design
strategy.
Here, analytical models to estimate the leakage current as a function of width has
been proposed. The derived equations depend on the structure (number of rows and
columns) of SRAM. Also the evaluation is based on the mode of operation (read,
write and idle phase).
2
Related Work
Since long time, static power estimation has been an area of research. Primarily, gate
level estimation was the focus but now, estimation at higher level is given more
attention. Butts and Sohi [2] proposed a generic model at micro-architectural level,
whose design was based on a key design parameter, k
design.
It captured device types,
geometries and stacking factors based on the simulations. Zhang et al. [8] develop an
architectural model for sub-threshold and gate leakage that captured temperature,
voltage, gate leakage, and parameter variations. Mamidipaka et al. developed
analytical models for subthreshold leakage estimation in memory arrays without spice
simulation [10].They also estimated width dependent leakage current models [1]. A
methodology for estimation of leakage power for micro-architectural components in
interconnection networks was proposed by Chen et al. [3]. The results were based on
the simulation of basic circuit components for varying input combination. Zhang et al.
proposed a model for leakage power of SRAM considering temperature, supply
voltage and bias voltage [9].
3
Proposed Work
In [1], width dependent leakage current of the SRAM was estimated in the form,
.
P is a constant
(3)
In our work, we have used the curve fitting tool of MATLAB, to obtain the linear
fitness equation for leakage current as a function of width. The equation, thus
obtained is in the form,
