Hardware Reference
In-Depth Information
With the encoding st1
D
00, st2
D
11, st3
D
01, st4
D
10. obtain the logic
equations for ns1.X; cs0; cs1/, ns0.X; cs0; cs1/ and Z.X; cs0; cs1/ (X input, Z
output, cs0; cs1 and ns0; ns1 present and next states).
Are there don't care minterms due to equivalent states ?
It is an open question of a good encoding that increases don't care minterms.
We leave also to reader the open question to characterize the sequential flexibility
beyond equivalent states.
14.2. Maximal sequential flexibility of a combinational node - continues
Given a specification network S, say that the fixed component F is all of S except
for a single node X, then the solution of the equation F
X
S gives the set of
all possible behaviors of X such that the functionality of S is maintained. This is
a superset of the MSF of X, since MSF may contain non-combinational behaviors,
whereas MSF is restricted to combinational ones.
Consider the following brute-force procedure to compute an approximation (i.e.,
maximal sequential flexibility) of the MSF of X (due to N.R. Satish):
1. Compute MSG
X
, the most general solution of node X, by solving the equation
F
X
S.
2. Set the flexibility to be CF plus the unreachable states of node X in the network.
3. Let M be the set of all care minterms of X.
4. For all minterms in M, replace in X the minterm by a don't care minterm to
obtain X
0
. Check sequential containment: if X
0
MSG
X
replace X by X
0
and
repeat with another care minterm.
What MSF is found depends on the ordering of minterms (trying all orderings would
give the MSF - maximum sequential flexibility).
(a) Apply the previous procedure to the following sequential circuit,
train11.
blif
:
.model train11.kiss2
.inputs v0 v1
.outputs v6.4
.latch
v6.0 v2
0
.latch
v6.1 v3
0
.latch
v6.2 v4
0
.latch
v6.3 v5
0
.names [20] v6.4
01
.names [12] v6.0
01
.names [14] v6.1
01
.names [16] v6.2
01
.names [18] v6.3
01
.names v1 v2 v3 v5 [0]
1000 1
.names v0 v4 v5 [1]
Search WWH ::
Custom Search