The name is absent

technology. As a result the circuit has a different power consumptions for each
input vector. Because of the process variation, the nominal power consumption
of the gate g_u is scaled by φ_u. For example, if input 1, input 2, and input 3 are
0, 1, and 1, respectively, then the total power consumption of the circuit would
be

POU = Pg₁,01Φl + Pg2,llΦ2 + Pg₃,0003 ÷ Pg₄,OO04

= 4.11201 + 15.1502 + 0.77603 + 17.410₄,          (3.2)

where p_3i,b> is the power consumption of the gate gi for input b). Note that 5J∙, the
input of each gate tj⅛, is a function of input vector of the circuit that is denoted
by bj. For example, in Figure 3.2, if bj = 011 then bɜ = 00.

In a digital circuit with N gates, for the binary input vector bj, total power
consumption pb_j is

N

Ph = ∑Pg_i⅛Φi-                      (3.3)

i=l

If there are M input vectors bɪ,..., Ьм, define measurement matrix A as

Pgi ,bj    Pg₃,bj ■ ■ ■ Pg_N Λ{^v

. Pgifi₃ Pg2J>l ∙ ∙ ∙ PgNfi₃
ʃl —                                      .

7⅛ι⅛ ⅛,⅛ ■ ∙ ■ Pgxfiu
Also, let

P = ∖Pb₁,Pb₂, ■ ■ ■ ,Pb_M\^T,

27

<<   33   34   35   36   37   38   39   >>

More intriguing information