I
n
t
e
r
n
at
ion
al
Jou
r
n
al
of
E
lec
t
r
ical
an
d
Com
p
u
t
e
r
E
n
gin
e
e
r
in
g
(
I
JE
CE
)
Vol.
15
,
No.
1
,
F
e
br
ua
r
y
20
25
,
pp.
1
63
~
1
7
3
I
S
S
N:
2088
-
8708
,
DO
I
:
10
.
11591/i
jec
e
.
v
15
i
1
.
pp
1
63
-
17
3
163
Jou
r
n
al
h
omepage
:
ht
tp:
//
ij
e
c
e
.
iaes
c
or
e
.
c
om
A
n
e
w
13
N
-
c
om
p
l
e
xi
t
y
m
e
m
o
r
y b
u
il
t
-
in
se
l
f
-
t
e
st
al
gor
it
h
m
t
o
b
al
an
c
e
st
at
i
c
r
a
n
d
om
a
c
c
e
ss
m
e
m
o
r
y st
at
i
c
f
au
l
t
c
ov
e
r
ag
e
an
d
t
e
st
t
i
m
e
Aim
an
Z
ak
wan
J
id
in
1
,
2
,
Raz
aid
i
Hu
s
s
i
n
3
,
M
oh
d
S
yaf
iq
M
is
p
an
1
,
2
,
L
e
e
We
n
g
F
ook
4
1
F
a
kul
ti
T
e
knol
ogi
da
n K
e
ju
r
ut
e
r
a
a
n E
le
kt
r
oni
k da
n K
omput
e
r
, U
ni
ve
r
s
it
i
T
e
kni
ka
l
M
a
la
ys
ia
M
e
la
ka
, D
ur
ia
n
T
ungga
l,
M
a
la
ys
ia
2
C
e
nt
e
r
f
or
T
e
le
c
omm
uni
c
a
ti
on R
e
s
e
a
r
c
h a
nd I
nnova
ti
on, Univ
e
r
s
it
i
T
e
kni
ka
l
M
a
la
ys
ia
M
e
la
ka
, D
ur
ia
n
T
ungga
l,
M
a
la
ys
ia
3
F
a
c
ul
ty
of
E
le
c
tr
oni
c
s
E
ngi
ne
e
r
in
g a
nd T
e
c
hnol
ogy, Unive
r
s
it
i
M
a
la
ys
ia
P
e
r
li
s
, A
r
a
u,
M
a
la
y
s
ia
4
E
me
r
a
ld
S
ys
te
m De
s
ig
n C
e
nt
e
r
, B
a
ya
n L
e
pa
s
, M
a
la
y
s
ia
Ar
t
icle
I
n
f
o
AB
S
T
RA
CT
A
r
ti
c
le
h
is
tor
y
:
R
e
c
e
ived
Apr
23,
2024
R
e
vis
e
d
Aug
27,
2024
Ac
c
e
pted
S
e
p
3,
2024
A
s
mem
o
ri
e
s
d
o
mi
n
at
e
t
h
e
s
y
s
t
em
-
on
-
c
h
i
p
(
So
C
)
,
t
h
e
i
r
q
u
a
l
i
t
y
s
i
g
n
i
fi
ca
n
t
l
y
i
mp
ac
t
s
t
h
e
c
h
i
p
ma
n
u
fac
t
u
r
i
n
g
y
i
el
d
.
T
h
ere
i
s
a
g
ro
w
i
n
g
n
ee
d
t
o
re
d
u
ce
t
h
e
ch
i
p
p
ro
d
u
ct
i
o
n
t
i
me
an
d
co
s
t
,
w
h
i
ch
mai
n
l
y
d
ep
e
n
d
s
o
n
t
h
e
t
e
s
t
i
n
g
p
h
as
e.
H
en
ce,
a
memo
r
y
b
u
i
l
t
-
i
n
s
e
l
f
-
t
e
s
t
(MBIS
T
)
u
t
i
l
i
zi
n
g
a
l
o
w
-
co
m
p
l
e
x
i
t
y
,
h
i
g
h
-
fa
u
l
t
-
co
v
erag
e
t
es
t
al
g
o
r
i
t
h
m
i
s
e
s
s
e
n
t
i
al
f
o
r
eff
i
ci
e
n
t
a
n
d
t
h
o
ro
u
g
h
memo
ry
t
e
s
t
i
n
g
.
T
h
e
March
A
Z
1
al
g
o
ri
t
h
m,
w
i
t
h
1
3
N
co
m
p
l
e
x
i
t
y
,
w
as
creat
ed
ear
l
i
er
t
o
b
a
l
an
ce
t
h
e
t
e
s
t
l
en
g
t
h
a
n
d
fa
u
l
t
co
v
e
rag
e.
H
o
w
e
v
er,
p
o
o
r
p
o
s
i
t
i
o
n
i
n
g
o
f
a
w
r
i
t
e
o
p
era
t
i
o
n
i
n
i
t
s
t
es
t
s
eq
u
en
ce
ca
u
s
ed
t
h
e
re
d
u
c
t
i
o
n
o
f
t
h
e
t
ra
n
s
i
t
i
o
n
co
u
p
l
i
n
g
fau
l
t
(CF
t
r)
d
et
ect
i
o
n
.
T
h
i
s
p
ap
er
p
re
s
en
t
s
t
h
e
creat
i
o
n
o
f
t
h
e
March
A
Z
al
g
o
ri
t
h
m,
mo
d
i
f
i
ed
fro
m
t
h
e
March
A
Z
1
al
g
o
ri
t
h
m,
t
o
i
n
creas
e
CFt
r
c
o
v
era
g
e
w
h
i
l
e
p
re
s
erv
i
n
g
t
h
e
s
ame
c
o
mp
l
ex
i
t
y
.
It
w
as
acc
o
mp
l
i
s
h
e
d
b
y
an
al
y
zi
n
g
t
h
e
fa
u
l
t
co
v
e
rag
e
o
ffered
b
y
t
h
e
Marc
h
A
Z
1
al
g
o
r
i
t
h
m
a
n
d
t
h
e
n
reo
rg
a
n
i
z
i
n
g
i
t
s
t
es
t
s
eq
u
e
n
ce
t
o
a
d
d
re
s
s
t
h
e
l
i
mi
t
a
t
i
o
n
i
n
d
et
ec
t
i
n
g
CF
t
r.
T
h
e
n
e
w
l
y
p
r
o
d
u
ce
d
Marc
h
A
Z
1
a
l
g
o
ri
t
h
m
w
as
s
u
cce
s
s
f
u
l
l
y
i
m
p
l
eme
n
t
e
d
i
n
an
MBIST
co
n
t
r
o
l
l
er.
T
h
e
s
i
mu
l
at
i
o
n
t
e
s
t
s
v
al
i
d
a
t
ed
i
t
s
f
u
n
c
t
i
o
n
a
l
i
t
y
a
n
d
d
emo
n
s
t
rat
e
d
t
h
at
t
h
e
CFt
r
co
v
era
g
e
w
a
s
en
h
a
n
ced
fro
m
6
2
.
5
%
t
o
7
5
%
,
ach
i
e
v
i
n
g
an
o
v
era
l
l
fau
l
t
co
v
era
g
e
o
f
8
3
.
3
%
.
T
h
eref
o
re,
w
i
t
h
1
3
N
co
mp
l
ex
i
t
y
,
i
t
o
ffers
t
h
e
b
es
t
fau
l
t
co
v
erag
e
amo
n
g
al
l
t
h
e
ex
i
s
t
i
n
g
t
es
t
al
g
o
ri
t
h
m
s
w
i
t
h
a
co
mp
l
ex
i
t
y
b
el
o
w
1
8
N
.
K
e
y
w
o
r
d
s
:
M
a
r
c
h
tes
t
a
lgor
it
hm
M
e
mor
y
buil
t
-
in
s
e
lf
-
tes
t
M
e
mor
y
f
a
ult
c
ove
r
a
ge
R
a
ndon
a
c
c
e
s
s
memor
y
Unlinked
s
tatic
f
a
ult
Th
i
s
i
s
a
n
o
p
en
a
c
ces
s
a
r
t
i
c
l
e
u
n
d
e
r
t
h
e
CC
B
Y
-
SA
l
i
ce
n
s
e.
C
or
r
e
s
pon
din
g
A
u
th
or
:
Aiman
Z
a
kwa
n
J
idi
n
F
a
kult
i
T
e
knologi
da
n
Ke
jur
ute
r
a
a
n
E
lektr
on
ik
da
n
Komputer
,
Unive
r
s
it
i
T
e
knikal
M
a
lays
ia
M
e
laka
76100
Dur
ian
T
ungga
l,
M
a
lays
ia
E
mail:
a
im
a
nz
a
kwa
n@utem.
e
du.
my
1.
I
NT
RODU
C
T
I
ON
M
e
mor
y
tes
ti
ng
is
be
c
omi
ng
e
s
s
e
nti
a
l
in
de
s
igni
ng
s
ys
tem
-
on
-
c
hips
(
S
oC
s
)
s
ince
they
a
r
e
nowa
da
ys
memor
y
domi
na
nt
,
whe
r
e
the
memor
ies
us
e
up
to
94%
of
their
a
r
e
a
s
[
1
]
–
[
3]
.
As
a
r
e
s
ult
,
a
go
od
c
hip
manuf
a
c
tur
ing
yield
is
s
igni
f
ica
ntl
y
in
f
luenc
e
d
by
memor
y
qua
li
ty
[
2]
,
[
4]
.
Additi
ona
ll
y
,
memo
r
ies
a
r
e
mor
e
pr
one
to
f
a
il
u
r
e
than
s
e
que
nti
a
l
logi
c
due
to
their
high
-
de
ns
it
y
na
tur
e
[
5]
.
M
a
ny
s
tatic
memor
y
f
a
ult
models
a
r
e
e
s
tablis
he
d
to
r
e
pr
e
s
e
nt
the
a
c
tual
manuf
a
c
tur
ing
de
f
e
c
t
a
t
the
logi
c
a
l
a
bs
tr
a
c
ti
on
leve
l,
a
s
de
s
c
r
ibed
in
T
a
ble
1.
S
tuck
-
a
t
f
a
ult
(
S
AF
)
,
tr
a
ns
it
ion
f
a
ult
(
T
F
)
,
r
e
a
d
de
s
tr
uc
ti
ve
f
a
ult
(
R
DF)
,
incor
r
e
c
t
r
e
a
d
f
a
ult
(
I
R
F
)
,
d
e
c
e
pti
ve
r
e
a
d
de
s
tr
uc
ti
ve
f
a
ult
(
DR
DF)
,
a
nd
wr
i
te
dis
tur
b
f
a
ult
(
W
DF)
a
r
e
c
las
s
if
ied
a
s
s
ingl
e
-
c
e
ll
f
a
ult
s
(
S
C
F
)
,
whos
e
oc
c
ur
r
e
nc
e
s
a
r
e
s
e
ns
it
ize
d
a
nd
de
tec
ted
in
the
s
a
me
memor
y
c
e
ll
.
M
e
a
nwhile,
t
r
a
ns
it
ion
c
oupli
ng
f
a
ult
(
C
F
t
r
)
,
de
c
e
pti
ve
r
e
a
d
de
s
tr
uc
ti
ve
c
oupli
ng
f
a
ult
(
C
F
dr
d
)
,
a
nd
wr
it
e
dis
tur
b
c
oupl
ing
f
a
ult
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2088
-
8708
I
nt
J
E
lec
&
C
omp
E
ng
,
Vol
.
15
,
No.
1
,
F
e
br
ua
r
y
20
25
:
1
63
-
1
7
3
164
(
C
F
wd)
a
r
e
the
double
-
c
e
ll
f
a
ult
s
(
DC
F
)
,
whe
r
e
th
e
f
a
ult
de
tec
ted
in
a
v
ictim
c
e
ll
(
v
)
is
c
a
us
e
d
by
th
e
s
tate
of
the
a
ggr
e
s
s
or
c
e
ll
(
a
)
[
6
]
–
[
8]
.
E
a
c
h
s
ingl
e
-
c
e
ll
f
a
ult
s
(
S
C
F
)
is
de
s
c
r
ibed
by
it
s
f
a
ult
s
pr
im
it
ives
(
F
P
)
a
nd
c
onve
nti
ona
ll
y
notate
d
a
s
<
/
/
>
,
whe
r
e
S
indi
c
a
tes
the
f
a
ult
s
e
ns
it
izing
ope
r
a
t
ion(
s
)
,
is
the
s
tate
i
f
f
a
ult
y,
a
nd
is
the
r
e
a
d
output
(
i
f
a
ppli
c
a
ble)
[
9]
,
[
10]
.
E
a
c
h
S
C
F
ha
s
2
F
P
s
s
ince
e
qua
ls
e
it
he
r
0
o
r
1
.
M
e
a
nwhile,
a
DC
F
’
s
F
P
is
notate
d
a
s
<
;
/
/
>
,
whe
r
e
a
nd
a
r
e
the
s
e
ns
it
izing
ope
r
a
tor
s
or
s
tate
s
a
t
the
a
nd
c
e
ll
s
,
r
e
s
pe
c
ti
ve
ly.
E
a
c
h
DC
F
c
ons
is
ts
of
8
s
ince
two
pos
s
ibl
e
s
c
e
na
r
ios
a
r
e
a
nti
c
ipate
d:
the
a
-
c
e
ll
’
s
a
ddr
e
s
s
is
inf
e
r
ior
(
<
)
or
s
upe
r
ior
(
>
)
to
the
v
-
c
e
ll
a
ddr
e
s
s
.
S
i
nc
e
numer
ous
memor
ies
on
a
c
hip
ne
e
d
to
be
tes
ted
a
utom
a
ti
c
a
ll
y,
memor
y
buil
t
-
in
s
e
lf
-
tes
t
(
M
B
I
S
T
)
is
a
wide
ly
us
e
d
method
f
o
r
memor
y
tes
ti
ng
[
11]
.
I
t
c
a
n
a
utom
a
te
tes
t
e
xe
c
uti
on
s
a
nd
output
c
he
c
king
,
a
nd
thus
,
the
de
pe
nde
nc
y
on
c
os
tl
y
tes
ti
ng
e
quipm
e
nt
is
r
e
duc
e
d
[
10]
,
[
12]
–
[
14
]
.
I
t
pe
r
f
o
r
ms
a
s
e
r
ies
of
tes
t
ope
r
a
ti
ons
de
f
ined
by
the
a
ppli
e
d
tes
t
a
lgor
it
hm,
c
ons
is
ti
ng
of
r
e
a
ding
(
)
or
wr
i
ti
ng
(
)
the
x
logi
c
to
e
ve
r
y
c
e
ll
ins
ide
the
tes
ted
memor
y
[
15]
,
[
16
]
.
T
he
s
e
tes
t
ope
r
a
ti
ons
a
r
e
c
onduc
ted
in
the
a
s
c
e
nding
(
⇑
)
or
de
s
c
e
nding
(
⇓
)
a
ddr
e
s
s
or
de
r
.
T
a
ble
1.
T
he
de
s
c
r
ipt
ions
o
f
unli
nke
d
s
tatic
f
a
ult
m
ode
ls
F
a
ul
t
FP
F
a
ul
ty
B
e
ha
vi
or
D
e
te
c
ti
on R
e
qui
r
e
me
nt
S
A
F
<
/
’
/
−
>
v
-
c
e
ll
i
s
s
tu
c
k a
t
th
e
x
-
s
ta
te
r
e
ga
r
dl
e
s
s
of
t
he
i
nput
va
lu
e
.
W
r
it
e
x’
to
c
e
ll
s
f
ol
lo
w
e
d by a
r
e
a
d ope
r
a
ti
on.
TF
<
’
/
/
−
>
v
-
c
e
ll
f
a
il
s
t
o t
r
a
ns
it
f
r
om
x
to
x
’.
W
r
it
e
x’
to
x
-
s
ta
te
c
e
ll
s
f
ol
lo
w
e
d
by a
r
e
a
d ope
r
a
ti
on.
R
D
F
<
/
’
/
’
>
A
r
e
a
d f
r
om t
he
v
-
c
e
ll
une
xpe
c
te
dl
y
c
ha
nge
s
i
ts
s
ta
te
a
nd r
e
tu
r
ns
a
n
in
c
or
r
e
c
t
va
lu
e
.
R
e
a
d f
r
om
x
-
s
ta
te
c
e
ll
s
.
I
R
F
<
/
/
’
>
A
r
e
a
d f
r
om t
he
v
-
c
e
ll
une
xpe
c
te
dl
y
r
e
tu
r
ns
a
n i
nc
or
r
e
c
t
va
lu
e
w
it
hout
c
ha
ngi
ng i
ts
s
ta
te
.
R
e
a
d f
r
om
x
-
s
ta
te
c
e
ll
s
.
D
R
D
F
<
/
’
/
>
A
r
e
a
d f
r
om t
he
v
-
c
e
ll
une
xpe
c
te
dl
y
c
ha
nge
s
i
ts
s
ta
te
but
r
e
tu
r
ns
t
he
c
or
r
e
c
t
va
lu
e
.
R
e
a
d t
w
ic
e
f
r
om
x
-
s
ta
te
c
e
ll
s
.
W
D
F
<
/
’
/
−
>
A
w
r
it
e
-
to
-
x
to
t
he
v
-
c
e
ll
t
ha
t
c
ont
a
in
s
a
n
x
une
xp
e
c
te
dl
y c
ha
ng
e
s
it
s
s
ta
te
t
o
x
’.
W
r
it
e
x
to
x
-
s
ta
te
c
e
ll
s
f
ol
lo
w
e
d
by a
r
e
a
d ope
r
a
ti
on.
C
F
tr
<
;
’
/
/
−
>
a
>
v
,
<
;
’
/
/
−
>
a
<
v
,
<
’
;
’
/
/
−
>
a
>
v
,
<
’
;
’
/
/
−
>
a
>
v
v
-
c
e
ll
f
a
il
s
t
o t
r
a
ns
it
f
r
om
x
to
x
’
w
he
n
its
a
-
c
e
ll
i
s
i
n a
gi
ve
n
s
ta
te
(
x
or
x
’).
W
r
it
e
x’
to
x
-
s
ta
te
c
e
ll
s
f
ol
lo
w
e
d
by a
r
e
a
d ope
r
a
ti
on w
he
n
a
-
c
e
ll
i
s
in
t
he
x
or
x’
s
ta
te
.
C
F
dr
d
<
;
/
’
/
>
a
>
v
,
<
;
/
’
/
>
a
<
v
,
<
’
;
/
’
/
>
a
>
v
,
<
’
;
/
’
/
>
a
<
v
A
r
e
a
d f
r
om t
he
v
-
c
e
ll
une
xpe
c
te
dl
y
c
ha
nge
s
i
ts
s
ta
te
but
r
e
tu
r
ns
t
he
c
or
r
e
c
t
va
lu
e
w
he
n i
ts
a
-
c
e
ll
i
s
i
n a
gi
ve
n s
ta
te
(
x
or
x
’).
R
e
a
d t
w
ic
e
f
r
om
x
-
s
ta
te
c
e
ll
s
w
he
n
a
-
c
e
ll
i
s
i
n t
he
x
or
x’
s
ta
te
.
C
F
w
d
<
;
/
’
/
−
>
a
>
v
,
<
;
/
’
/
−
>
a
<
v
,
<
’
;
/
’
/
−
>
a
>
v
,
<
’
;
/
’
/
−
>
a
>
v
A
w
r
it
e
-
to
-
x
to
t
he
v
-
c
e
ll
t
ha
t
c
ont
a
in
s
a
n
x
une
xp
e
c
te
dl
y c
ha
ng
e
s
it
s
s
ta
te
t
o
x
’
w
he
n i
ts
a
-
c
e
ll
i
s
i
n a
gi
ve
n s
ta
te
(
x
or
x
’).
W
r
it
e
x
to
x
-
s
ta
te
c
e
ll
s
f
ol
lo
w
e
d
by a
r
e
a
d ope
r
a
ti
on w
he
n
a
-
c
e
ll
i
s
in
t
he
x
or
x’
s
ta
te
.
T
he
s
e
mi
c
onduc
tor
indus
tr
y
p
r
e
f
e
r
s
M
a
r
c
h
tes
t
a
lgor
it
hms
s
ince
they
ha
ve
de
s
ign
s
im
pli
c
it
y
a
nd
li
ne
a
r
c
ompl
e
xit
y,
de
f
ined
in
the
or
de
r
of
N
(
the
s
i
z
e
of
the
tes
ted
memor
y)
[
2]
,
[
17]
–
[
19]
.
S
e
ve
r
a
l
M
a
r
c
h
tes
t
a
lgor
it
hms
a
r
e
li
s
ted
in
T
a
ble
2
.
T
he
y
a
r
e
dis
ti
nguis
he
d
by
their
tes
t
s
e
que
nc
e
s
,
c
ompl
e
xit
ies
,
a
nd
f
a
ult
c
ove
r
a
ge
s
.
T
he
s
tuck
-
a
t
f
a
ult
(
S
AF)
r
e
p
r
e
s
e
nts
incor
r
e
c
t
r
e
a
d
f
a
ult
(
I
R
F
)
a
nd
r
e
a
d
de
s
tr
uc
ti
ve
f
a
ul
t
(
R
DF)
c
ove
r
a
ge
s
s
ince
their
de
tec
ti
on
r
e
quir
e
ments
a
r
e
a
li
ke
[
20]
.
T
he
s
hown
f
a
ult
c
ove
r
a
ge
is
c
omp
uted
by
divi
ding
the
number
of
de
tec
table
F
P
s
by
2
f
or
e
a
c
h
S
C
F
a
nd
by
8
f
o
r
e
a
c
h
DC
F
.
A
M
a
r
c
h
tes
t
a
lgor
it
hm
with
a
c
ompl
e
xit
y
higher
than
or
e
qua
l
to
18N,
li
k
e
the
M
a
r
c
h
M
S
S
a
lgor
it
hm
[
20]
,
o
f
f
e
r
s
c
ompl
e
te
c
ove
r
a
ge
of
a
ll
tar
ge
ted
s
tatic
f
a
ult
s
in
s
tatic
r
a
ndom
a
c
c
e
s
s
memor
y
(
S
R
AM
)
.
M
e
a
nwhile,
a
lowe
r
-
c
ompl
e
xit
y
tes
t
a
lgor
it
hm
is
ne
c
e
s
s
a
r
y
to
pr
oduc
e
a
s
hor
ter
tes
t
ti
me
a
nd
lowe
r
c
os
t.
How
e
ve
r
,
ba
s
e
d
on
T
a
ble
2,
it
ha
s
poor
c
ove
r
a
ge
of
DR
DF,
W
DF,
C
F
dr
d,
a
nd
C
F
wd,
whi
c
h
a
r
e
r
e
leva
nt
to
memo
r
ies
f
a
br
ica
ted
us
ing
the
n
a
nomete
r
pr
oc
e
s
s
tec
hnologi
e
s
[
21]
.
T
he
r
e
f
or
e
,
the
M
a
r
c
h
A
Z
1
(
13N)
a
nd
M
a
r
c
h
AZ
2
(
14N)
a
lgor
it
hms
we
r
e
c
r
e
a
ted
to
ba
lanc
e
the
c
ompl
e
xit
y
a
nd
c
ove
r
a
ge
of
the
tar
ge
ted
f
a
ult
s
[
22]
.
T
he
f
o
r
mer
o
f
f
e
r
s
80
.
6%
o
f
ove
r
a
ll
f
a
ult
c
ove
r
a
ge
,
pr
ovidi
ng
c
ompl
e
te
S
C
F
c
ove
r
a
ge
,
62
.
5%
c
ove
r
a
ge
of
C
F
tr
,
a
nd
75%
c
ove
r
a
ge
of
C
F
dr
d
a
n
d
C
F
wd.
M
e
a
nwhile,
the
latter
of
f
e
r
s
a
s
li
ght
e
nha
nc
e
ment
in
C
F
tr
c
ove
r
a
ge
(
75%
)
,
thus
of
f
e
r
ing
83.
3
%
o
f
ove
r
a
ll
f
a
ult
c
ove
r
a
ge
,
the
be
s
t
a
mong
a
ll
e
xis
ti
ng
be
low
18N
-
c
ompl
e
xit
y
tes
t
a
lgor
it
hms
[
22]
.
T
he
latter
c
a
n
de
tec
t
a
s
pe
c
if
ic
F
P
of
C
F
tr
(
C
F
tr
<
1
;
1
0
/
1
/
−
>
a
>
v
)
that
is
unde
tec
table
by
the
f
or
me
r
.
How
e
ve
r
,
it
s
c
omp
lexity
is
1N
mor
e
than
the
f
o
r
mer
,
r
e
qui
r
ing
a
s
li
ghtl
y
long
e
r
tes
t
ti
me.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
E
lec
&
C
omp
E
ng
I
S
S
N:
2088
-
8708
A
ne
w
13N
-
c
omple
x
it
y
me
mor
y
buil
t
-
in
s
e
lf
-
tes
t
al
gor
it
hm
to
balance
…
(
A
iman
Z
ak
w
an
J
idi
n
)
165
T
his
pa
pe
r
p
r
e
s
e
nts
the
M
a
r
c
h
A
Z
a
lgor
it
hm,
a
n
e
w
tes
t
a
lgor
it
hm
that
im
pr
ove
s
the
M
a
r
c
h
AZ
1
a
lgor
it
hm’
s
c
ove
r
a
ge
of
C
F
tr
whi
le
maintaining
it
s
c
ompl
e
xit
y
a
t
13N.
I
t
wa
s
a
c
c
ompl
is
he
d
by
a
na
lyzing
the
de
tec
tabili
ty
of
a
ll
F
P
s
us
ing
a
n
a
utom
a
ted
f
a
ult
de
tec
ti
on
a
na
lyze
r
,
whic
h
identif
ies
e
a
c
h
F
P
’
s
s
e
ns
it
izing
a
nd
de
tec
ti
ng
tes
t
ope
r
a
ti
ons
with
in
the
M
a
r
c
h
AZ
1
a
lgor
it
hm
’
s
tes
t
s
e
que
nc
e
.
S
ubs
e
que
ntl
y,
the
we
a
kne
s
s
in
C
F
tr
de
tec
ti
on
wa
s
r
e
c
ognize
d
f
r
om
the
a
na
lys
is
output
a
nd
a
ddr
e
s
s
e
d
thr
ough
tes
t
ope
r
a
ti
ons
a
nd
tes
t
e
leme
nts
r
e
or
ga
niza
ti
on.
T
he
f
unc
ti
ona
li
ty
of
the
ne
w
M
a
r
c
h
AZ
a
lgor
it
hm
wa
s
ve
r
if
ied
via
a
s
im
ulation
c
onduc
ted
us
ing
the
im
pleme
nted
M
B
I
S
T
c
ontr
oll
e
r
.
F
inally,
it
s
f
a
ult
c
ove
r
a
ge
wa
s
e
va
luate
d
by
pe
r
f
or
mi
ng
a
tes
t
on
a
f
a
ult
-
inj
e
c
ted
S
R
AM
a
s
the
memor
y
mo
de
l
in
the
s
im
ulation.
T
he
r
e
s
ult
s
de
mons
tr
a
te
that
the
ne
w
M
a
r
c
h
AZ
a
lgor
it
hm
p
r
ovides
s
im
il
a
r
unli
nke
d
s
tatic
f
a
ult
c
ove
r
a
ge
to
the
M
a
r
c
h
AZ
2
a
lgo
r
it
hm
,
whic
h
of
f
e
r
s
the
be
s
t
c
ove
r
a
ge
to
da
te
a
mong
a
ll
e
xis
ti
n
g
tes
t
a
lgo
r
it
hms
wi
th
a
c
ompl
e
xit
y
lowe
r
than
1
8N
[
22]
.
How
e
ve
r
,
with
1N
c
ompl
e
xit
y
les
s
e
r
than
the
latter
,
the
f
or
me
r
pr
oduc
e
s
a
f
a
s
ter
tes
t
c
ompl
e
ti
on
ti
me
a
nd,
thus
,
c
a
n
r
e
duc
e
the
tes
t
c
os
t.
T
a
ble
2.
S
e
ve
r
a
l
M
a
r
c
h
a
lgo
r
it
hms
tes
t
s
e
que
nc
e
s
,
c
ompl
e
xit
ies
,
a
nd
f
a
ult
c
ove
r
a
ge
s
T
e
s
t
a
lg
or
it
hm
C
ompl
e
xi
ty
T
e
s
t
s
e
que
nc
e
S
C
F
D
C
F
S
A
F
TF
D
R
D
F
W
D
F
C
F
tr
C
F
dr
d
C
F
w
d
M
a
r
c
h
C
-
[
6]
10N
⇕
(
w
0)
;
⇑
(
r
0, w
1)
;
⇑
(
r
1, w
0)
;
⇓
(
r
0, w
1)
;
⇓
(
r
1, w
0)
;
⇕
(
r
0
)
100%
100%
0%
0%
100%
0%
0%
M
a
r
c
h
C
L
[
23]
12N
⇕
(
w
0)
;
⇑
(
r
0, w
1)
;
⇑
(
r
1, r
1, w
0)
;
⇓
(
r
0,
w
1, r
1)
;
⇓
(
r
1, w
0)
;
⇕
(
r
0)
100%
100%
50%
0%
100%
50%
0%
M
a
r
c
h
L
R
[
24]
14N
⇕
(
w
0)
;
⇓
(
r
0, w
1)
;
⇑
(
r
1, w
0, r
0, w
1)
;
⇑
(
r
1, w
0)
;
⇑
(
r
0, w
1,
r
1, w
0)
;
⇑
(
r
0)
100%
100%
0%
0%
100%
0%
0%
M
a
r
c
h
S
R
[
6]
14N
⇕
(
w
0)
;
⇑
(
r
0, w
1, r
1, w
0
)
;
⇑
(
r
0, r
0)
;
⇑
(
w
1)
;
⇓
(
r
1, w
0, r
0, w
1
)
;
⇓
(
r
1, r
1)
100%
100%
100%
0%
100%
50%
0%
M
a
r
c
h
C
+
[
25]
14N
⇕
(
w
0)
;
⇑
(
r
0, w
1, r
1
)
;
⇑
(
r
1, w
0, r
0)
;
⇓
(
r
0, w
1, r
1
)
;
⇓
(
r
1, w
0, r
0)
;
⇕
(
r
0)
100%
100%
100%
0%
100%
100%
0%
M
a
r
c
h
A
Z
1
[
22]
13N
⇕
(
w
0)
;
⇓
(
r
0, w
1)
;
⇑
(
w
1, r
1,
r
1, w
0)
;
⇑
(
w
0, r
0)
;
⇑
(
r
0, w
1, w
1,
r
1)
;
⇑
(
r
1)
100%
100%
100%
100%
62.5%
75%
75%
M
a
r
c
h
A
Z
2
[
22]
14N
⇕
(
w
0)
;
⇓
(
w
0, r
0)
;
⇑
(
r
0, w
1, w
1, r
1)
;
⇑
(
r
1, w
0)
;
⇓
(
r
0, w
1, w
1,
r
1)
;
⇑
(
r
1)
;
100%
100%
100%
100%
75%
75%
75%
M
a
r
c
h
M
S
S
[
20]
18N
⇕
(
w
0)
;
⇑
(
r
0, r
0, w
1, w
1
)
;
⇑
(
r
1, r
1, w
0,
w
0)
;
⇓
(
r
0,
r
0, w
1, w
1)
;
⇓
(
r
1, r
1, w
0,
w
0)
;
⇕
(
r
0
)
100%
100%
100%
100%
100%
100%
100%
M
a
r
c
h
SS
[
2]
22N
⇕
(
w
0)
;
⇑
(
r
0, r
0, w
1, w
1
)
;
⇑
(
r
1, r
1, w
0,
w
0)
;
⇓
(
r
0,
r
0, w
1, w
1)
;
⇓
(
r
1, r
1, w
0,
w
0)
;
⇕
(
r
0
)
100%
100%
100%
100%
100%
100%
100%
2.
T
HE
M
AR
CH
AZ
1
AL
GO
RI
T
HM
RE
V
I
E
W
T
a
ble
3
s
hows
the
s
ix
tes
t
e
l
e
ments
in
the
M
a
r
c
h
AZ
1
a
lgor
it
hm’
s
tes
t
s
e
que
nc
e
,
labe
ll
e
d
0
thr
ough
5
,
s
e
pa
r
a
ted
by
s
e
mi
c
olons
[
22
]
.
T
he
tes
t
e
leme
nts
will
be
e
xe
c
uted
s
e
que
nti
a
ll
y
dur
ing
the
tes
t:
All
tes
t
ope
r
a
ti
ons
de
f
ined
in
mus
t
be
pe
r
f
or
me
d
on
a
ll
memo
r
y
c
e
ll
s
be
f
or
e
movi
ng
on
to
the
ne
xt
+
1
.
P
lus
,
13
r
e
a
d
or
wr
it
e
ope
r
a
ti
ons
mus
t
be
pe
r
f
or
med
on
a
ll
N
memo
r
y
c
e
ll
s
,
e
xplaining
it
s
1
3
N
c
ompl
e
xit
y.
T
a
ble
3.
T
he
M
a
r
c
h
A
Z
1
a
lgo
r
it
hm
de
s
c
r
ipt
ions
T
e
s
t
e
l
e
me
nt
T
e
s
t
s
e
que
nc
e
T
e
s
t
de
s
c
r
ip
ti
on
TE
0
⇕
(
w
0)
A
ll
c
e
ll
s
a
r
e
s
e
t
to
0.
TE
1
⇓
(
w
1)
A
ll
c
e
ll
s
a
r
e
s
e
t
to
1 i
n de
s
c
e
ndi
ng a
ddr
e
s
s
or
de
r
.
TE
2
⇑
(
w
1, r
1, r
1, w
0)
A
ll
c
e
ll
s
a
r
e
s
e
qu
e
nt
ia
ll
y s
e
t
to
1, r
e
a
d t
w
ic
e
(
e
xp
e
c
ti
ng a
1 a
t
t
he
out
put
)
, a
nd s
e
t
to
0 i
n a
s
c
e
ndi
ng a
ddr
e
s
s
or
de
r
.
TE
3
⇑
(
w
0, r
0)
A
ll
c
e
ll
s
a
r
e
s
e
qu
e
nt
ia
ll
y s
e
t
to
0 be
f
or
e
be
in
g r
e
a
d (
e
xpe
c
ti
ng a
0 a
t
th
e
out
put
)
i
n a
s
c
e
ndi
ng a
ddr
e
s
s
or
de
r
.
TE
4
⇑
(
r
0, w
1, w
1, r
1)
A
ll
c
e
ll
s
a
r
e
s
e
qu
e
nt
ia
ll
y r
e
a
d (
e
xpe
c
ti
ng 0)
, s
e
t
to
1 t
w
ic
e
,
a
nd
r
e
r
e
a
d
(
e
xpe
c
ti
ng 1)
i
n a
s
c
e
ndi
ng a
ddr
e
s
s
or
de
r
.
TE
5
⇑
(
r
1)
A
ll
c
e
ll
s
a
r
e
r
e
a
d (
e
xpe
c
ti
ng 1)
i
n a
s
c
e
ndi
ng a
ddr
e
s
s
or
de
r
.
A
f
a
ult
de
tec
ti
on
a
na
lys
is
wa
s
c
ondu
c
ted
on
the
M
a
r
c
h
AZ
1
a
lgor
it
hm
us
ing
a
de
ve
loped
f
a
ult
de
tec
ti
on
a
na
lyze
r
that
identif
ies
the
s
e
ns
it
ize
r
a
nd
de
tec
tor
pa
ir
s
f
o
r
a
ll
tar
ge
ted
F
P
s
wi
thi
n
the
tes
t
s
e
que
nc
e
[
26]
.
T
he
f
lowc
ha
r
t
in
F
igu
r
e
1
de
picts
the
a
na
lys
is
pr
oc
e
s
s
that
wa
s
c
onduc
ted.
Onc
e
the
M
a
r
c
h
AZ
1
a
lgor
it
hm’
s
tes
t
s
e
que
nc
e
wa
s
r
e
a
d
a
nd
e
xtr
a
c
ted,
the
a
na
lyze
r
de
ter
mi
ne
d
the
c
e
ll
tr
e
nd
o
f
e
a
c
h
tes
t
e
leme
nt,
whic
h
indi
c
a
tes
how
the
c
e
ll
s
’
s
t
a
tes
a
r
e
c
ha
ng
e
d
whe
n
a
tes
t
e
leme
nt
is
e
xe
c
uted
dur
ing
the
tes
t.
Ne
xt,
it
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2088
-
8708
I
nt
J
E
lec
&
C
omp
E
ng
,
Vol
.
15
,
No.
1
,
F
e
br
ua
r
y
20
25
:
1
63
-
1
7
3
166
identif
ied
a
ll
pos
s
ibl
e
s
e
ns
it
ize
r
a
nd
de
tec
tor
pa
i
r
s
of
e
a
c
h
de
tec
table
F
P
f
ound
withi
n
the
a
na
ly
z
e
d
tes
t
s
e
que
nc
e
,
s
tar
ti
ng
f
r
om
the
f
i
r
s
t
tes
t
ope
r
a
ti
on
de
f
ined
in
0
unt
il
the
las
t
tes
t
ope
r
a
ti
on
in
5
,
ba
s
e
d
o
n
their
de
tec
ti
on
r
e
qui
r
e
ments
de
s
c
r
ibed
in
T
a
ble
1.
T
he
p
r
oc
e
s
s
wa
s
r
e
pe
a
ted
f
or
a
ll
36
tar
ge
t
e
d
F
P
s
.
S
pe
c
if
ica
ll
y,
f
o
r
DC
F
de
tec
ti
on
a
na
lys
is
,
the
pr
e
d
e
ter
mi
ne
d
c
e
ll
tr
e
nds
,
whic
h
indi
c
a
te
the
wa
y
a
ll
memor
y
c
e
ll
s
’
c
ontents
c
ha
nge
dur
ing
the
e
xe
c
uti
on
of
a
tes
t
e
leme
nt,
a
r
e
ne
e
de
d
to
de
c
ide
the
c
or
r
e
s
po
nding
F
P
(
e
it
he
r
<
or
>
)
[
26]
.
F
igur
e
1.
T
he
f
a
ult
de
tec
ti
on
a
na
lys
is
pr
oc
e
s
s
f
low
E
a
c
h
F
P
wa
s
a
s
s
oc
iate
d
with
a
bit
in
the
de
t_F
P
b
us
f
or
f
a
ult
c
ove
r
a
ge
c
omput
a
ti
on
pu
r
pos
e
s
,
whic
h
wa
s
s
e
t
to
high
whe
n
it
s
s
e
ns
it
ize
r
-
de
tec
tor
pa
ir
w
a
s
identif
ied
withi
n
the
a
na
lyze
d
tes
t
s
e
que
nc
e
.
T
he
r
e
f
or
e
,
the
f
a
ult
c
ove
r
a
ge
wa
s
c
omput
e
d
by
c
a
lcul
a
ti
ng
the
high
de
t_F
P
bit
s
divi
de
d
by
the
tot
a
l
F
P
s
36.
T
a
ble
4
s
hows
the
s
e
ns
it
ize
r
a
nd
de
tec
tor
pa
ir
s
f
or
e
a
c
h
F
P
identif
ied
with
in
the
M
a
r
c
h
A
Z
1
a
lgo
r
it
hm’
s
tes
t
s
e
que
nc
e
dur
ing
the
a
na
lys
is
.
T
he
TE
i
-
j
notations
s
igni
f
y
tha
t
the
j
th
tes
t
ope
r
a
ti
on
in
T
E
i
is
r
e
c
ognize
d
a
s
a
s
e
ns
it
izing
or
de
tec
ti
ng
ope
r
a
ti
on
f
or
a
pa
r
ti
c
ular
F
P
.
T
a
ble
4
de
mons
tr
a
tes
that
a
ll
tar
ge
ted
S
C
F
s
a
r
e
de
t
e
c
table
s
ince
their
F
P
s
ha
ve
a
t
lea
s
t
one
identi
f
ied
s
e
ns
it
ize
r
-
d
e
tec
tor
pa
ir
.
S
o,
the
M
a
r
c
h
AZ
1
a
lg
or
it
hm
of
f
e
r
s
100%
of
a
ll
S
C
F
s
.
Additi
ona
ll
y
,
the
f
a
ult
a
na
lyze
r
identif
ied
the
s
e
ns
it
ize
r
-
de
tec
tor
pa
ir
s
f
or
5
F
P
s
of
C
F
tr
.
He
nc
e
,
C
F
t
r
c
ove
r
a
ge
e
qua
ls
62.
5
%
(
5
de
tec
table
F
P
s
out
of
8)
.
M
e
a
nwhile,
the
f
a
ult
a
na
lyze
r
identif
ied
the
s
e
ns
it
ize
r
-
de
tec
tor
pa
ir
s
f
or
6
F
P
s
o
f
e
a
c
h
C
F
dr
d
a
nd
C
F
wd.
He
nc
e
,
the
C
F
d
r
d
a
nd
C
F
wd
c
ove
r
a
ge
s
e
qua
l
75%
(
6
de
tec
table
F
P
s
o
ut
of
8)
.
C
ons
e
que
ntl
y,
the
f
a
ult
de
tec
ti
on
a
na
lys
is
de
r
ived
the
e
xpe
c
ted
f
a
ult
c
ove
r
a
ge
by
the
M
a
r
c
h
AZ
1
a
l
gor
it
hm,
a
s
pr
e
s
e
nted
in
T
a
ble
2.
B
y
c
ompar
ing
it
s
f
a
ul
t
c
ove
r
a
ge
to
the
M
a
r
c
h
AZ
2
a
lgor
it
h
m
with
1
4N
tes
t
c
ompl
e
xit
y
[
22]
,
whic
h
is
a
va
il
a
ble
in
T
a
ble
2,
the
a
na
lyze
d
M
a
r
c
h
AZ
1
a
lgor
it
hm
ha
s
a
s
li
ghtl
y
lowe
r
c
ove
r
a
ge
of
C
F
tr
s
ince
it
c
a
nnot
de
tec
t
the
C
F
t
r
<
1
;
1
0
/
1
/
−
>
a
>
v
,
a
s
p
r
ove
n
by
the
a
na
lys
is
output
pr
e
s
e
nted
in
T
a
ble
4
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
E
lec
&
C
omp
E
ng
I
S
S
N:
2088
-
8708
A
ne
w
13N
-
c
omple
x
it
y
me
mor
y
buil
t
-
in
s
e
lf
-
tes
t
al
gor
it
hm
to
balance
…
(
A
iman
Z
ak
w
an
J
idi
n
)
167
T
a
ble
4.
T
he
a
na
lys
is
of
the
M
a
r
c
h
AZ
1
a
lgor
it
h
m
’
s
f
a
ult
c
ove
r
a
ge
F
a
ul
t
FP
I
de
nt
if
ie
d (
S
e
ns
it
iz
e
r
, D
e
te
c
to
r
)
D
e
te
c
ti
on
s
ta
tu
s
F
a
ul
t
c
ove
r
a
ge
S
A
F
<
1/
0/
-
>
(
T
E
2
-
1
, T
E
2
-
2
)
, (
T
E
4
-
3
,
T
E
4
-
4
)
Y
e
s
2/
2 (
100%
)
<
0/
1/
-
>
(
T
E
3
-
1
, T
E
3
-
2
)
Y
e
s
TF
<
0w
1/
0/
-
>
(
T
E
1
-
1
, T
E
2
-
2
)
, (
T
E
4
-
2
,
T
E
4
-
4
)
Y
e
s
2/
2
(
100%
)
<
1w
0/
1/
-
>
(
T
E
2
-
4
, T
E
3
-
2
)
Y
e
s
R
D
F
<
r
0/
1/
1>
(
T
E
2
-
1
, T
E
2
-
2
)
, (
T
E
4
-
3
,
T
E
4
-
4
)
Y
e
s
2/
2 (
100%
)
<
r
1/
0/
0>
(
T
E
3
-
1
, T
E
3
-
2
)
Y
e
s
I
R
F
<
r
0/
0/
1>
(
T
E
2
-
1
, T
E
2
-
2
)
, (
T
E
4
-
3
,
T
E
4
-
4
)
Y
e
s
2/
2 (
100%
)
<
r
1/
1/
0>
(
T
E
3
-
1
, T
E
3
-
2
)
Y
e
s
D
R
D
F
<
r
0/
1/
0>
(
T
E
3
-
2
, T
E
4
-
1
)
Y
e
s
2/
2 (
100%
)
<
r
1/
0/
1>
(
T
E
2
-
2
, T
E
2
-
3
)
, (
T
E
4
-
4
,
T
E
5
-
1
)
Y
e
s
W
D
F
<
0w
0/
1/
-
>
(
T
E
3
-
1
, T
E
3
-
2
)
Y
e
s
2/
2 (
100%
)
<
1w
1/
0/
-
>
(
T
E
2
-
1
, T
E
2
-
2
)
, (
T
E
4
-
3
,
T
E
4
-
4
)
Y
e
s
C
F
tr
<
0;
0w
1/
0/
-
>
a
>
v
(
T
E
4
-
2
, T
E
4
-
4
)
Y
e
s
5/
8 (
62.5%
)
<
0;
0w
1/
0/
-
>
a
<
v
(
T
E
1
-
1
, T
E
2
-
2
)
Y
e
s
<
1;
0w
1/
0/
-
>
a
>
v
(
T
E
1
-
1
, T
E
2
-
2
)
Y
e
s
<
1;
0w
1/
0/
-
>
a
<
v
(
T
E
4
-
2
, T
E
4
-
4
)
Y
e
s
<
0;
1w
0/
1/
-
>
a
>
v
N
ot
f
ound
No
<
0;
1w
0/
1/
-
>
a
<
v
(
T
E
2
-
4
, T
E
3
-
2
)
Y
e
s
<
1;
1w
0/
1/
-
>
a
>
v
N
ot
f
ound
No
<
1;
1w
0/
1/
-
>
a
<
v
N
ot
f
ound
No
C
F
dr
d
<
0;
r
0/
1/
0>
a
>
v
(
T
E
3
-
2
, T
E
4
-
1
)
Y
e
s
6/
8
(
75%
)
<
0;
r
0/
1/
0>
a
<
v
(
T
E
3
-
2
, T
E
4
-
1
)
Y
e
s
<
1;
r
0/
1/
0>
a
>
v
N
ot
f
ound
No
<
1;
r
0/
1/
0>
a
<
v
N
ot
f
ound
No
<
0;
r
1/
0/
1>
a
>
v
(
T
E
4
-
4
, T
E
5
-
1
)
Y
e
s
<
0;
r
1/
0/
1>
a
<
v
(
T
E
2
-
2
, T
E
2
-
3
)
Y
e
s
<
1;
r
1/
0/
1>
a
>
v
(
T
E
2
-
2
, T
E
2
-
3
)
Y
e
s
<
1;
r
1/
0/
1>
a
<
v
(
T
E
4
-
4
, T
E
5
-
1
)
Y
e
s
C
F
w
d
<
0;
0w
0/
1/
-
>
a
>
v
(
T
E
3
-
1
, T
E
3
-
2
)
Y
e
s
6/
8 (
75%
)
<
0;
0w
0/
1/
-
>
a
<
v
(
T
E
3
-
1
, T
E
3
-
2
)
Y
e
s
<
1;
0w
0 /
1/
-
>
a
>
v
N
ot
f
ound
No
<
1;
0w
0/
1/
-
>
a
<
v
N
ot
f
ound
No
<
0;
1w
1/
0/
-
>
a
>
v
(
T
E
4
-
3
, T
E
4
-
4
)
Y
e
s
<
0;
1w
1/
0/
-
>
a
<
v
(
T
E
2
-
1
, T
E
2
-
2
)
Y
e
s
<
1;
1w
1/
0/
-
>
a
>
v
(
T
E
2
-
1
,
TE
2
-
2
)
Y
e
s
<
1;
1w
1/
0/
-
>
a
<
v
(
T
E
4
-
3
, T
E
4
-
4
)
Y
e
s
3.
T
HE
NE
W
M
AR
CH
AZ
AL
GO
RI
T
HM
CR
E
A
T
I
ON
As
s
tate
d
in
T
a
ble
1
,
a
C
F
tr
oc
c
ur
r
e
nc
e
c
a
n
be
s
e
ns
it
ize
d
in
a
v
-
c
e
ll
by
wr
it
ing
a
n
x’
logi
c
to
the
c
e
ll
that
c
ontains
a
n
x
logi
c
whe
n
the
a
-
c
e
ll
is
in
a
g
i
ve
n
s
tate
.
T
he
n,
the
wr
it
e
ope
r
a
ti
on
is
s
uc
c
e
e
de
d
by
a
r
e
a
d
ope
r
a
ti
on
to
de
tec
t
a
ny
f
a
ult
y
be
ha
vior
f
r
om
the
v
-
c
e
ll
.
Ac
c
or
ding
to
[
20]
,
[
22]
,
the
C
F
tr
<
1;
1w0/1/
-
>
a
>
v
c
a
n
be
s
e
ns
it
ize
d
a
nd
de
tec
ted
by
us
ing
one
of
the
f
oll
owing
tes
t
s
e
que
nc
e
s
,
whe
r
e
F
(
x
)
r
e
pr
e
s
e
nts
a
ny
o
pe
r
a
ti
on
that
pr
oduc
e
s
a
n
x
-
s
tate
in
the
memor
y
c
e
ll
s
a
nd
*
indi
c
a
tes
that
the
a
s
s
oc
iate
d
ope
r
a
ti
ons
a
r
e
opti
ona
l:
−
C
ondit
ion
3.
1:
⇕
(
…
,
(
1
)
)
;
⇑
(
(
1
)
∗
,
0
,
0
∗
0
,
(
0
)
∗
)
;
−
C
ondit
ion
3.
2:
⇕
(
…
,
(
1
)
)
;
⇑
(
(
1
)
∗
,
0
,
0
∗
)
;
⇕
(
0
,
…
)
;
−
C
ondit
ion
3.
3:
⇕
(
…
,
(
1
)
)
;
⇕
(
0
,
0
∗
,
0
,
(
0
)
∗
,
(
1
)
)
;
I
n
the
M
a
r
c
h
AZ
1
a
lgor
it
hm
’
s
tes
t
s
e
que
nc
e
,
the
c
e
ll
s
’
tr
a
ns
it
ion
f
r
om
1
to
0
c
a
n
only
oc
c
ur
a
t
TE
2
:
⇑
(
1
,
1
,
1
,
0
)
,
whe
r
e
the
0
ope
r
a
ti
on
s
hould
s
e
t
the
c
e
ll
s
’
s
tate
s
to
0.
A
s
ubs
e
que
nt
r
e
a
d
ope
r
a
ti
on
c
a
n
then
de
tec
t
the
f
a
ult
y
be
ha
viour
c
a
us
e
d
by
the
C
F
tr
<
1;
1w0/1/
-
>
a
>
v
.
Ye
t
,
thi
s
0
ope
r
a
ti
on
in
TE
2
is
f
oll
owe
d
by
a
nother
w0
ope
r
a
ti
on
in
TE
3
:
⇑
(
0
,
0
)
be
f
o
r
e
the
r
e
quir
e
d
r
e
a
d
ope
r
a
ti
on.
T
he
r
e
f
or
e
,
thi
s
tes
t
s
e
que
nc
e
doe
s
not
mee
t
C
ondit
ion
3.
1
to
C
ondit
io
n
3.
3
r
e
qui
r
e
ments
.
I
n
f
a
c
t,
the
0
ope
r
a
ti
on
in
TE
3
a
c
ts
a
s
the
C
F
tr
<
1;
1w0/1/
-
>
a
>
v
f
a
ult
r
e
c
ove
r
e
d
,
mas
king
it
s
oc
c
ur
r
e
nc
e
f
r
om
be
ing
de
tec
ted
by
the
0
ope
r
a
t
ion
in
TE
3
,
a
s
il
lus
tr
a
ted
in
F
igur
e
2
us
ing
a
4
-
c
e
ll
me
mor
y
a
s
the
e
xa
mpl
e
whe
r
e
the
v
-
c
e
ll
a
nd
a
-
c
e
ll
a
r
e
s
e
t
to
a
ddr
e
s
s
0
a
nd
2,
r
e
s
pe
c
ti
ve
ly.
I
n
TE
2
o
pe
r
a
ti
on
,
th
e
v
-
c
e
ll
,
a
f
f
e
c
ted
by
the
C
F
t
r
<
1;
1w0/
1/
-
>
a
>
v
f
a
ult
,
f
a
i
ls
to
c
ha
nge
i
ts
s
tate
t
o
low
w
he
n
t
he
w
0
ope
r
a
t
ion
is
pe
r
f
o
r
me
d
s
ince
it
s
a
-
c
e
ll
(
c
e
l
l
2
)
is
in
a
high
s
tate
.
S
omehow
,
the
w
0
ope
r
a
ti
on
in
TE
3
s
uc
c
e
s
s
f
ul
ly
c
ha
nge
s
i
ts
s
ta
te
to
low
s
inc
e
it
s
a
-
c
e
ll
is
no
l
onge
r
in
a
hi
gh
s
tate
.
S
o,
the
M
a
r
c
h
AZ
1
a
lgor
it
hm
’
s
TE
2
a
nd
TE
3
we
r
e
r
e
or
ga
nize
d
to
s
olve
thi
s
is
s
ue
:
the
0
ope
r
a
ti
on
in
TE
3
wa
s
moved
to
the
e
nd
o
f
TE
2
.
S
ubs
e
que
ntl
y,
the
ne
wly
modi
f
ied
TE
2
c
ons
is
ts
of
⇑
(
w
1,
r
1,
r
1,
w
0,
w
0)
tes
t
s
e
que
nc
e
,
whe
r
e
a
s
the
ne
w
tes
t
s
e
que
nc
e
f
or
TE
3
is
⇑
(
0
)
.
C
ons
e
que
ntl
y,
the
ne
wly
r
e
or
ga
nize
d
TE
2
a
nd
TE
3
f
ul
f
il
the
r
e
qui
r
e
d
tes
t
s
e
que
nc
e
de
f
ined
b
y
C
ondit
ion
3
.
2
a
nd
s
hould
be
a
ble
to
de
tec
t
the
C
F
tr
<
1;
1w0/1/
-
>
a
>
v
.
T
he
ne
wly
modi
f
ied
M
a
r
c
h
AZ
1
a
lgor
it
hm
is
c
a
ll
e
d
the
M
a
r
c
h
AZ
a
lgor
it
hm
,
with
t
he
s
a
me
13N
c
ompl
e
xit
y
a
nd
the
ne
w
tes
t
s
e
que
nc
e
:
⇕
(
0
)
;
⇓
(
0
,
1
)
;
⇑
(
1
,
1
,
1
,
0
,
0
)
;
⇑
(
0
)
;
⇑
(
0
,
1
,
1
,
1
)
;
⇑
(
1
)
.
T
he
f
a
ult
de
tec
ti
on
a
na
lys
is
wa
s
r
e
done
us
ing
the
ne
w
M
a
r
c
h
AZ
a
lgor
i
thm
.
T
he
a
na
lys
is
r
e
s
ult
s
in
T
a
ble
5
pr
ove
that
the
C
F
tr
c
ove
r
a
ge
wa
s
im
pr
o
ve
d
f
r
om
62
.
5%
by
the
M
a
r
c
h
A
Z
1,
a
s
s
tate
d
in
T
a
ble
4,
to
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2088
-
8708
I
nt
J
E
lec
&
C
omp
E
ng
,
Vol
.
15
,
No.
1
,
F
e
br
ua
r
y
20
25
:
1
63
-
1
7
3
168
75%
by
e
na
bli
ng
the
de
tec
ti
on
of
the
C
F
tr
<
1;
1w0/1/
-
>
a
>
v
.
F
ur
ther
mo
r
e
,
it
a
ls
o
pr
ove
s
that
the
pr
op
os
e
d
tes
t
e
leme
nt
r
e
or
ga
niza
ti
on
did
not
a
f
f
e
c
t
the
de
tec
ti
on
s
of
other
F
P
s
a
s
their
c
ove
r
a
ge
s
r
e
main
unc
ha
nge
d.
F
igur
e
2.
I
l
lus
tr
a
ti
on
of
the
C
F
t
r
<
1;
1w0/1/
-
>
a
>
v
f
a
ult
r
e
c
ove
r
ing
a
t
T
E
3
of
the
M
a
r
c
h
A
Z
1
a
lgor
it
h
m
T
a
ble
5.
T
he
ne
w
M
a
r
c
h
AZ
a
lgor
it
hm’
s
f
a
ult
de
te
c
ti
on
a
na
lys
is
F
a
ul
t
FP
I
de
nt
if
ie
d (
S
e
ns
it
iz
e
r
, D
e
te
c
to
r
)
D
e
te
c
ti
on s
ta
tu
s
F
a
ul
t
c
ove
r
a
ge
S
A
F
<
1/
0/
-
>
(
T
E
2
-
1
, T
E
2
-
2
)
, (
T
E
4
-
3
,
T
E
4
-
4
)
Y
e
s
2/
2 (
100%
)
<
0/
1/
-
>
(
T
E
2
-
5
, T
E
3
-
1
)
Y
e
s
TF
<
0w
1/
0/
-
>
(
T
E
1
-
1
, T
E
2
-
2
)
, (
T
E
4
-
2
,
T
E
4
-
4
)
Y
e
s
2/
2 (
100%
)
<
1w
0/
1/
-
>
(
T
E
2
-
4
, T
E
3
-
1
)
Y
e
s
R
D
F
<
r
0/
1/
1>
(
T
E
2
-
1
, T
E
2
-
2
)
, (
T
E
4
-
3
,
T
E
4
-
4
)
Y
e
s
2/
2 (
100%
)
<
r
1/
0/
0>
(
T
E
2
-
5
, T
E
3
-
1
)
Y
e
s
I
R
F
<
r
0/
0/
1>
(
T
E
2
-
1
, T
E
2
-
2
)
, (
T
E
4
-
3
,
T
E
4
-
4
)
Y
e
s
2/
2 (
100%
)
<
r
1/
1/
0>
(
T
E
2
-
5
, T
E
3
-
1
)
Y
e
s
D
R
D
F
<
r
0/
1/
0>
(
T
E
3
-
1
, T
E
4
-
1
)
Y
e
s
2/
2 (
100%
)
<
r
1/
0/
1>
(
T
E
2
-
2
, T
E
2
-
3
)
, (
T
E
4
-
4
,
T
E
5
-
1
)
Y
e
s
W
D
F
<
0w
0/
1/
-
>
(
T
E
2
-
5
, T
E
3
-
1
)
Y
e
s
2/
2 (
100%
)
<
1w
1/
0/
-
>
(
T
E
2
-
1
, T
E
2
-
2
)
, (
T
E
4
-
3
,
T
E
4
-
4
)
Y
e
s
C
F
tr
<
0;
0w
1/
0/
-
>
a
>
v
(
T
E
4
-
2
, T
E
4
-
4
)
Y
e
s
6/
8 (
75%
)
<
0;
0w
1/
0/
-
>
a
<
v
(
T
E
1
-
1
, T
E
2
-
2
)
Y
e
s
<
1;
0w
1/
0/
-
>
a
>
v
(
T
E
1
-
1
, T
E
2
-
2
)
Y
e
s
<
1;
0w
1/
0/
-
>
a
<
v
(
T
E
4
-
2
, T
E
4
-
4
)
Y
e
s
<
0;
1w
0/
1/
-
>
a
>
v
N
ot
f
ound
No
<
0;
1w
0/
1/
-
>
a
<
v
(
T
E
2
-
4
, T
E
3
-
1
)
Y
e
s
<
1;
1w
0/
1/
-
>
a
>
v
(
T
E
2
-
4
, T
E
3
-
1
)
Y
e
s
<
1;
1w
0/
1/
-
>
a
<
v
N
ot
f
ound
No
C
F
dr
d
<
0;
r
0/
1/
0>
a
>
v
(
T
E
3
-
1
, T
E
4
-
1
)
Y
e
s
6/
8 (
75%
)
<
0;
r
0/
1/
0>
a
<
v
(
T
E
3
-
1
, T
E
4
-
1
)
Y
e
s
<
1;
r
0/
1/
0>
a
>
v
N
ot
f
ound
No
<
1;
r
0/
1/
0>
a
<
v
N
ot
f
ound
No
<
0;
r
1/
0/
1>
a
>
v
(
T
E
4
-
4
, T
E
5
-
1
)
Y
e
s
<
0;
r
1/
0/
1>
a
<
v
(
T
E
2
-
2
, T
E
2
-
3
)
Y
e
s
<
1;
r
1/
0/
1>
a
>
v
(
T
E
2
-
2
, T
E
2
-
3
)
Y
e
s
<
1;
r
1/
0/
1>
a
<
v
(
T
E
4
-
4
, T
E
5
-
1
)
Y
e
s
C
F
w
d
<
0;
0w
0/
1/
-
>
a
>
v
(
T
E
2
-
5
, T
E
3
-
1
)
Y
e
s
6/
8 (
75%
)
<
0;
0w
0/
1/
-
>
a
<
v
(
T
E
2
-
5
, T
E
3
-
1
)
Y
e
s
<
1;
0w
0/
1/
-
>
a
>
v
N
ot
f
ound
No
<
1;
0w
0/
1/
-
>
a
<
v
N
ot
f
ound
No
<
0;
1w
1/
0/
-
>
a
>
v
(
T
E
4
-
3
, T
E
4
-
4
)
Y
e
s
<
0;
1w
1/
0/
-
>
a
<
v
(
T
E
2
-
1
, T
E
2
-
2
)
Y
e
s
<
1;
1w
1/
0/
-
>
a
>
v
(
T
E
2
-
1
, T
E
2
-
2
)
Y
e
s
<
1;
1w
1/
0/
-
>
a
<
v
(
T
E
4
-
3
, T
E
4
-
4
)
Y
e
s
4.
RE
S
UL
T
S
AN
D
DI
S
CU
S
S
I
ON
T
he
ne
w
M
a
r
c
h
AZ
a
lgor
i
thm
’
s
tes
t
s
e
que
nc
e
wa
s
ha
r
d
-
c
ode
d
a
s
the
us
e
r
-
de
f
ined
a
lgor
it
hm
(
UD
A)
ins
ide
a
n
M
B
I
S
T
c
ontr
oll
e
r
,
ge
ne
r
a
ted
us
ing
S
ieme
ns
T
e
s
s
e
nt
memor
y
B
I
S
T
s
of
twa
r
e
a
s
the
e
lec
tr
onic
de
s
ign
a
utom
a
ti
on
(
E
DA
)
tool
.
Af
ter
that
,
it
wa
s
s
im
ulate
d
in
the
s
ieme
ns
Que
s
taSi
m
s
im
ulator
us
ing
the
c
r
e
a
ted
tes
t
be
nc
he
s
a
nd
tes
t
pa
tt
e
r
ns
.
T
wo
dif
f
e
r
e
nt
tes
ts
we
r
e
c
onduc
ted
in
s
im
ulations
on
the
im
pl
e
mente
d
M
B
I
S
T
c
ontr
oll
e
r
:
a
tes
t
on
a
f
a
ult
-
f
r
e
e
memor
y
a
nd
a
tes
t
on
a
f
a
ult
-
inj
e
c
ted
memor
y
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
E
lec
&
C
omp
E
ng
I
S
S
N:
2088
-
8708
A
ne
w
13N
-
c
omple
x
it
y
me
mor
y
buil
t
-
in
s
e
lf
-
tes
t
al
gor
it
hm
to
balance
…
(
A
iman
Z
ak
w
an
J
idi
n
)
169
4.
1.
T
e
s
t
on
t
h
e
f
au
lt
-
f
r
e
e
m
e
m
or
y
m
od
e
l
T
his
tes
t
a
s
s
e
s
s
e
d
the
f
unc
ti
ona
li
ty
of
the
c
r
e
a
ted
M
B
I
S
T
with
the
ne
w
M
a
r
c
h
AZ
a
lgor
it
hm
a
s
the
UD
A.
I
t
wa
s
e
va
luate
d
by
obs
e
r
ving
the
M
B
I
ST
P
G_G
O
f
lag,
whic
h
s
hould
s
tay
high
unti
l
the
t
e
s
t
wa
s
c
ompl
e
ted
or
whe
n
the
M
B
I
ST
P
G_D
ON
E
f
lag
w
a
s
a
s
s
e
r
ted.
T
he
tes
t
c
ompl
e
ti
on
t
im
e
wa
s
a
ls
o
m
e
a
s
ur
e
d,
whic
h
s
hould
e
qua
l
the
UD
A’
s
c
ompl
e
xit
y
mul
ti
pli
e
d
by
N
a
nd
the
c
lock
pe
r
iod
(
20
ns
)
.
A
1
-
kB
memor
y
wa
s
us
e
d
a
s
the
te
s
t
memor
y;
thus
,
N
e
qua
ls
1024
.
F
igur
e
3
pr
e
s
e
nts
the
s
im
ulation
wa
ve
f
or
m
in
Q
ue
s
taSi
m
f
or
the
f
a
ult
-
f
r
e
e
memor
y
tes
t.
T
he
outpu
t
da
ta
r
e
a
d
f
r
om
the
memor
y
c
e
ll
(
dout
)
wa
s
c
ompar
e
d
to
the
e
xpe
c
ted
da
ta
ge
ne
r
a
ted
by
the
M
B
I
S
T
c
ontr
oll
e
r
(
B
I
ST
_
E
X
P
E
C
T
_DA
T
A
)
whe
ne
ve
r
C
M
P
_E
N
is
h
igh.
T
he
M
B
I
ST
P
G_G
O
f
lag
wa
s
a
s
s
e
r
ted
to
indi
c
a
te
the
s
tar
t
of
the
tes
t
a
nd
r
e
maine
d
high
unt
il
the
tes
t
c
o
mpl
e
ti
on,
a
s
s
igni
f
ied
by
a
high
M
B
I
ST
P
G_D
ON
E
f
lag.
T
h
is
obs
e
r
va
ti
on
s
igni
f
ies
no
dis
c
r
e
p
a
nc
y
be
twe
e
n
dout
a
nd
B
I
ST
_E
X
P
E
C
T
_DA
T
A
dur
ing
the
c
ompar
is
on.
A
ddit
ionally,
the
tes
t
c
ompl
e
ti
on
ti
me
,
mea
s
ur
e
d
f
r
om
the
s
tar
t
unti
l
the
e
nd
o
f
the
tes
t,
e
qua
ls
266
,
240
ns
.
I
t
is
s
im
il
a
r
to
the
e
xpe
c
ted
tes
t
c
ompl
e
ti
on
ti
me
s
ince
13*1024*20
ns
e
qua
ls
266
,
240
ns
.
T
he
r
e
f
or
e
,
th
is
tes
t’
s
obs
e
r
va
ti
on
va
li
da
ted
the
i
mpl
e
mente
d
M
B
I
S
T
’
s
c
or
r
e
c
t
f
unc
ti
ona
li
ty,
whic
h
us
e
d
the
M
a
r
c
h
AZ
a
s
the
UD
A
.
F
ur
ther
mo
r
e
,
it
a
ls
o
de
mons
tr
a
tes
that
the
ne
w
M
a
r
c
h
AZ
a
lgor
it
hm
pr
oduc
e
s
a
tes
t
20
,
480
ns
s
hor
ter
than
the
M
a
r
c
h
AZ
2
a
lgor
it
hm
,
whic
h
r
e
quir
e
s
286
,
720
ns
[
22
]
,
on
the
s
a
me
memor
y
model
unde
r
tes
t.
F
igur
e
3.
T
he
s
im
ulation
wa
ve
f
or
m
obs
e
r
ve
d
in
Que
s
taSi
m
f
r
om
the
tes
t
on
the
f
a
ult
-
f
r
e
e
memor
y
m
ode
l
4.
2.
T
e
s
t
on
t
h
e
f
au
lt
-
in
j
e
c
t
e
d
m
e
m
or
y
m
od
e
l
T
his
tes
t
a
s
s
e
s
s
e
d
the
f
a
u
lt
c
o
ve
r
a
g
e
of
the
a
ppl
ied
M
a
r
c
h
A
Z
a
lg
or
it
h
m.
T
he
be
ha
v
io
r
a
l
m
ode
l
of
the
memo
r
y
us
e
d
in
t
he
pr
e
v
ious
tes
t
wa
s
mo
di
f
ied
t
o
i
nt
r
oduc
e
a
l
l
F
P
s
t
o
b
e
d
e
tec
te
d
a
nd
,
he
nc
e
,
im
i
tate
the
ir
f
a
ul
ty
be
ha
vio
r
s
du
r
i
ng
t
he
tes
t
in
the
s
i
mul
a
ti
o
n.
F
igu
r
e
4
s
hows
the
dis
t
r
i
but
ion
o
f
the
a
f
f
e
c
ted
v
ic
ti
m
c
e
ll
s
f
or
a
ll
in
tr
oduc
e
d
F
P
s
(
n
otate
d
a
s
V
i
)
a
nd
the
c
o
r
r
e
s
pondin
g
a
gg
r
e
s
s
or
c
e
l
ls
(
no
tate
d
a
s
A
i
)
f
or
e
a
c
h
DC
F
;
the
a
ddr
e
s
s
e
s
o
f
thes
e
c
e
ll
s
we
r
e
r
a
ndo
ml
y
ge
ne
r
a
te
d
withi
n
the
give
n
s
pe
c
if
i
c
a
ti
ons
,
e
.
g
.
,
th
e
a
d
dr
e
s
s
of
A
i
m
us
t
be
g
r
e
a
te
r
t
ha
n
V
i
f
or
a
ll
F
P
s
wi
th
a
>
v
.
T
he
S
A
F
<
x
/x’
/
-
>
wa
s
int
r
od
uc
e
d
by
f
ix
ing
the
in
put
da
ta
di
n
va
lue
to
x
whe
n
t
he
w
r
i
te
-
e
na
ble
s
ig
na
l
we
wa
s
hi
gh
a
nd
th
e
wa
s
e
qu
a
l
to
the
a
f
f
e
c
ted
c
e
ll
c
h
os
e
n.
T
he
R
DF
<r
x
/
x’
/
x’
>
oc
c
u
r
r
e
nc
e
s
we
r
e
in
tr
oduc
e
d
b
y
a
lt
e
r
in
g
the
low
s
tate
o
f
the
a
f
f
e
c
ted
v
-
c
e
ll
s
to
h
igh
whe
n
th
e
y
we
r
e
a
bout
to
be
r
e
a
d
(
w
e
is
l
ow)
.
I
n
c
o
ntr
a
s
t,
t
he
I
R
F
<
r
x
/
x
/
x’
>
oc
c
ur
r
e
nc
e
s
we
r
e
r
e
pl
ica
te
d
b
y
o
ve
r
w
r
i
ti
ng
the
outpu
t
v
a
lue
t
o
x’
whe
n
the
a
f
f
e
c
te
d
v
-
c
e
ll
s
c
onta
in
e
d
l
ogic
x
a
nd
we
r
e
a
bo
ut
to
be
r
e
a
d
.
T
he
C
F
tr
<
y
;
x
w
x’
/
x
/
-
>
a
>
v
a
nd
C
F
tr
<
x
;
x
w
x’
/
x
/
-
>
a
>
v
oc
c
ur
r
e
nc
e
s
we
r
e
pr
oduc
e
d
by
c
a
nc
e
ll
ing
the
w
x’
ope
r
a
ti
on
on
to
the
a
f
f
e
c
ted
v
-
c
e
ll
s
that
c
ontai
ne
d
x
whe
n
the
c
or
r
e
s
ponding
a
-
c
e
ll
s
a
r
e
in
y
-
s
tat
e
,
whe
r
e
y
=
{0,
1}.
At
the
s
a
me
ti
me
,
the
T
F
<
x
w
x’
/
x
/
-
>
is
c
ons
ider
e
d
de
tec
table
whe
n
a
t
lea
s
t
one
C
F
tr
<
y
;
x
w
x’
/
x
/
-
>
wa
s
de
tec
ted.
M
e
a
nwhile,
the
oc
c
ur
r
e
nc
e
s
of
C
F
d
r
d
<
y
;
r
x
/
x’
/
x
>
a
>
v
a
nd
C
F
d
r
d
<
y
;
r
x
/
x’
/
x
>
a
>
v
we
r
e
im
it
a
ted
by
a
lt
e
r
ing
the
c
ontents
o
f
the
a
f
f
e
c
ted
v
-
c
e
ll
s
c
ontaining
logi
c
x
to
x’
whe
n
they
we
r
e
r
e
a
d
a
nd
the
c
or
r
e
s
ponding
a
-
c
e
ll
s
c
ontaine
d
logi
c
y
.
S
im
il
a
r
l
y,
DR
DF
<
r
x
/
x’
/
x
>
is
c
ons
ider
e
d
de
tec
table
whe
n
a
t
lea
s
t
one
C
F
dr
d
<
y
;
r
x
/
x’
/
x
>
wa
s
de
tec
ted.
L
a
s
tl
y,
the
oc
c
ur
r
e
nc
e
s
of
C
F
wd
<
y
;
x
w
x
/
x’
/
-
>
a
>
v
a
nd
C
F
wd
<
y
;
x
w
x
/
x’
/
-
>
a
<
v
we
r
e
c
r
e
a
ted
by
c
ha
nging
the
inp
ut
din
va
lue
to
x’
whe
n
the
a
f
f
e
c
ted
v
-
c
e
ll
s
c
ontai
ne
d
logi
c
x
a
nd
we
r
e
a
bout
to
be
r
e
wr
i
tt
e
n
to
x
,
a
nd
whe
n
the
c
or
r
e
s
ponding
a
-
c
e
ll
s
s
tor
e
d
logi
c
y
.
He
nc
e
,
W
DF
<
x
w
x
/
x’
/
-
>
is
de
e
med
de
tec
table
whe
n
a
t
lea
s
t
one
C
F
wd
<
y
;
x
w
x
/
x’
/
-
>
wa
s
de
tec
ted.
F
igur
e
5
d
is
plays
the
s
im
ulation
wa
ve
f
o
r
m
of
the
M
B
I
S
T
ope
r
a
ti
on
on
the
f
a
ult
-
inj
e
c
ted
memor
y
us
ing
the
ne
w
M
a
r
c
h
AZ
a
s
the
UD
A.
I
n
thi
s
tes
t,
the
va
lues
of
a
ll
f
a
ult
de
tec
ti
on
f
lags
we
r
e
obs
e
r
v
e
d
whe
n
the
tes
t
wa
s
c
ompl
e
ted
(
indi
c
a
ted
by
a
high
M
B
I
S
T
P
G_D
ON
E
f
lag
)
a
nd
r
e
c
or
de
d
in
T
a
ble
6
.
T
he
r
e
f
or
e
,
the
M
a
r
c
h
AZ
a
lgor
it
hm
’
s
f
a
ult
c
ove
r
a
ge
wa
s
de
ter
mi
ne
d
by
c
ounti
ng
the
high
b
it
s
in
e
a
c
h
f
a
ult
’
s
de
tec
t
ion
f
lag.
I
t
s
hows
that
the
M
a
r
c
h
AZ
a
lgor
it
hm
de
tec
ted
the
inj
e
c
ted
C
F
tr
<
1;
1w0/1/
-
>
a
>
v
,
whic
h
wa
s
unde
tec
ted
by
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2088
-
8708
I
nt
J
E
lec
&
C
omp
E
ng
,
Vol
.
15
,
No.
1
,
F
e
br
ua
r
y
20
25
:
1
63
-
1
7
3
170
the
M
a
r
c
h
AZ
1
a
lgo
r
it
hm.
T
he
r
e
f
or
e
,
the
f
or
mer
pr
ovides
a
be
tt
e
r
C
F
tr
(
75%
)
a
nd
ove
r
a
ll
f
a
ult
c
ove
r
a
ge
(
83.
3%
)
than
the
latter
(
62
.
5%
a
nd
80
.
6%
,
r
e
s
pe
c
ti
ve
ly)
.
I
t
pr
ovides
s
im
il
a
r
f
a
ult
c
ove
r
a
ge
c
ompar
e
d
to
the
14N
-
c
ompl
e
xit
y
M
a
r
c
h
AZ
2
a
lgor
it
hm,
whos
e
f
a
ult
c
ove
r
a
ge
is
pr
e
s
e
nted
in
T
a
ble
2.
How
e
ve
r
,
s
ince
it
s
c
ompl
e
xit
y
is
1N
lowe
r
than
the
M
a
r
c
h
AZ
2
a
lgor
it
hm,
it
s
M
B
I
S
T
ope
r
a
ti
on
may
r
e
quir
e
a
s
hor
ter
c
ompl
e
ti
on
ti
me.
C
ons
e
que
ntl
y,
a
s
pr
ove
n
by
the
s
im
ulation
r
e
s
ult
s
obtaine
d
f
r
om
the
tes
ts
on
both
f
a
ult
-
f
r
e
e
a
nd
f
a
ult
-
inj
e
c
ted
memor
ies
,
the
ne
w
M
a
r
c
h
AZ
a
lgor
it
hm,
with
13N
c
ompl
e
xit
y
,
of
f
e
r
s
the
be
s
t
ba
lanc
e
be
twe
e
n
memor
y
tes
ti
ng
ti
me
a
nd
f
a
ult
c
ove
r
a
ge
s
ince
it
pr
ovides
the
be
s
t
c
ove
r
a
ge
of
the
tar
ge
t
e
d
f
a
ult
s
a
mong
a
ll
e
xis
ti
ng
tes
t
a
lgor
it
hms
with
a
c
ompl
e
xit
y
be
low
18N
a
nd
pr
oduc
e
s
a
s
hor
ter
tes
t
ti
me
than
the
M
a
r
c
h
AZ
2
a
lgor
it
hm
.
F
igur
e
4.
T
he
dis
tr
ibut
ion
of
the
a
f
f
e
c
ted
v
-
c
e
ll
s
a
nd
the
c
or
r
e
s
ponding
a
-
c
e
ll
s
(
f
o
r
DC
F
)
in
the
f
a
ult
-
i
njec
ted
memor
y
model
us
e
d
f
or
the
s
im
ulation
F
i
gu
r
e
5
.
T
h
e
wa
ve
f
or
m
obs
e
r
ve
d
f
r
om
th
e
t
e
s
t
on
th
e
f
a
u
l
t
-
i
nj
e
c
t
e
d
me
mo
r
y
,
us
in
g
t
he
M
a
r
c
h
A
Z
a
s
the
UD
A
Evaluation Warning : The document was created with Spire.PDF for Python.
I
nt
J
E
lec
&
C
omp
E
ng
I
S
S
N:
2088
-
8708
A
ne
w
13N
-
c
omple
x
it
y
me
mor
y
buil
t
-
in
s
e
lf
-
tes
t
al
gor
it
hm
to
balance
…
(
A
iman
Z
ak
w
an
J
idi
n
)
171
T
a
ble
6.
T
he
M
a
r
c
h
A
Z
a
lgo
r
it
hm’
s
f
a
ult
c
ove
r
a
ge
de
r
ived
f
r
om
the
s
im
ulation
F
a
ul
t
D
e
te
c
ti
on F
la
g
O
bs
e
r
ve
d D
e
te
c
ti
on F
la
g V
a
lu
e
D
e
r
iv
e
d F
a
ul
t
C
ove
r
a
ge
S
A
F
s
a
f
_de
t
11
b
2 de
te
c
te
d F
P
s
out
of
2 (
100%
)
TF
tf
_de
t
11
b
2 de
te
c
te
d F
P
s
out
of
2 (
100%
)
R
D
F
r
df
_de
t
11
b
2 de
te
c
te
d F
P
s
out
of
2 (
100%
)
I
R
F
ir
f
_de
t
11
b
2 de
te
c
te
d F
P
s
out
of
2 (
100%
)
D
R
D
F
dr
df
_de
t
11
b
2 de
te
c
te
d F
P
s
out
of
2
(
100%
)
W
D
F
w
df
_de
t
11
b
2 de
te
c
te
d F
P
s
out
of
2 (
100%
)
C
F
tr
c
f
tr
_de
t
11110110
b
6 de
te
c
te
d F
P
s
out
of
8 (
75%
)
C
F
dr
d
c
f
dr
d_de
t
11001111
b
6 de
te
c
te
d F
P
s
out
of
8 (
75%
)
C
F
w
d
c
f
w
d_de
t
11001111
b
6 de
te
c
te
d F
P
s
out
of
8 (
75%
)
O
ve
r
a
ll
f
a
ul
t
c
ove
r
a
ge
30
de
te
c
te
d F
P
s
out
of
36 (
83.3%
)
5.
CONC
L
USI
ON
T
his
a
r
ti
c
le
int
r
oduc
e
s
the
ne
w
mar
c
h
AZ
a
lgo
r
i
thm
,
whic
h
e
nha
nc
e
s
the
C
F
tr
a
nd
ove
r
a
ll
f
a
ult
c
ove
r
a
ge
s
of
f
e
r
e
d
by
the
e
xis
ti
ng
M
a
r
c
h
AZ
1
a
lgor
it
hm
while
ke
e
ping
the
c
ompl
e
xit
y
a
t
13N.
T
h
e
M
a
r
c
h
AZ
1
a
lgor
i
thm
wa
s
f
ir
s
t
a
na
lyze
d
to
identi
f
y
it
s
w
e
a
kne
s
s
in
de
tec
ti
ng
C
F
tr
due
to
the
poo
r
pos
it
ioni
ng
of
a
wr
it
e
ope
r
a
ti
on
.
He
nc
e
,
two
tes
t
e
leme
nts
withi
n
it
s
tes
t
s
e
que
nc
e
we
r
e
r
e
or
ga
nize
d
by
movi
ng
the
id
e
nti
f
ied
w0
ope
r
a
ti
on
f
r
o
m
one
tes
t
e
leme
nt
to
a
nother
t
o
s
olve
the
unne
c
e
s
s
a
r
y
C
F
tr
r
e
c
ove
r
y
a
nd
im
pr
ove
C
F
tr
de
tec
ti
on
without
invol
ving
a
ddit
ional
tes
t
ope
r
a
ti
ons
.
T
he
ne
wly
c
r
e
a
ted
tes
t
s
e
que
nc
e
,
the
M
a
r
c
h
AZ
a
lgor
it
hm,
wa
s
r
e
a
na
lyze
d
to
e
ns
ur
e
that
C
F
tr
c
ove
r
a
ge
wa
s
im
pr
ove
d
without
a
f
f
e
c
ti
ng
oth
e
r
f
a
ult
s
’
de
tec
ti
ons
.
I
t
then
s
e
r
ve
d
a
s
the
tes
t
a
lgor
it
hm
in
t
he
im
pleme
nted
M
B
I
S
T
c
ontr
oll
e
r
,
whic
h
wa
s
late
r
us
e
d
in
s
im
ulations
to
c
onduc
t
tes
ts
on
two
d
if
f
e
r
e
nt
memor
y
models
.
T
he
f
ir
s
t
tes
t
on
a
f
a
ult
-
f
r
e
e
memor
y
de
mons
tr
a
ted
it
s
c
or
r
e
c
t
f
unc
ti
ona
li
ty,
a
s
no
mi
s
m
a
tch
be
twe
e
n
the
r
e
a
d
a
nd
e
xpe
c
ted
da
ta
wa
s
f
oun
d
dur
ing
the
s
im
ulation.
T
he
n,
the
s
e
c
ond
tes
t
c
onduc
ted
on
a
f
a
ult
-
inj
e
c
ted
memor
y
va
li
da
ted
that
,
w
it
h
13N
c
ompl
e
xit
y,
i
t
o
f
f
e
r
s
83.
3%
o
f
ove
r
a
ll
f
a
ult
c
ove
r
a
ge
,
s
im
il
a
r
to
the
f
a
ult
c
ove
r
a
ge
p
r
ovided
by
th
e
M
a
r
c
h
AZ
2
a
lgor
it
hm
with
14N
c
ompl
e
xit
y
.
C
ons
e
que
ntl
y,
it
o
f
f
e
r
s
the
be
s
t
ba
lanc
e
be
twe
e
n
the
tes
t
c
o
mpl
e
ti
on
ti
me
a
nd
f
a
ult
c
ove
r
a
ge
a
mong
a
ll
tes
t
a
lgo
r
it
hm
s
with
a
c
ompl
e
xit
y
lowe
r
than
18N
s
ince
it
pr
o
vides
the
highes
t
c
ove
r
a
ge
of
the
int
e
nde
d
f
a
ult
s
with
only
1
3N
tes
t
c
ompl
e
xit
y.
AC
KNOWL
E
DGE
M
E
NT
S
T
he
a
uthor
s
would
li
ke
to
a
c
knowle
dge
Unive
r
s
it
i
T
e
knikal
M
a
lays
ia
M
e
laka
(
UT
e
M
)
,
Unive
r
s
it
i
M
a
lays
ia
P
e
r
li
s
(
UniM
AP)
,
a
nd
the
M
ini
s
tr
y
of
Hi
ghe
r
E
duc
a
ti
on
M
a
lays
ia
f
or
their
c
ont
r
ibut
ion,
a
s
s
is
tanc
e
,
a
nd
s
uppor
t
to
thi
s
r
e
s
e
a
r
c
h
unde
r
the
r
e
s
e
a
r
c
h
gr
a
nt
F
R
GS/
1/2024/
T
K07/U
T
E
M
/02/
17.
RE
F
E
RE
NC
E
S
[
1]
N
is
ha
O
.
S.
a
nd
S
iv
a
s
a
nk
a
r
K
.,
“
A
r
c
hi
te
c
tu
r
e
f
or
a
n
e
f
f
ic
ie
nt
M
B
I
S
T
us
in
g
modi
f
ie
d
ma
r
c
h
-
y
a
lg
or
it
hms
to
a
c
hi
e
ve
opt
im
i
z
e
d
c
omm
uni
c
a
ti
on
de
la
y
a
nd
c
omput
a
ti
ona
l
s
p
e
e
d,”
I
nt
e
r
nat
io
na
l
J
our
nal
of
P
e
r
v
as
iv
e
C
om
put
in
g
and
C
om
m
uni
c
at
io
n
s
,
vol
.
17,
no. 1, pp. 135
–
147, F
e
b. 2021, doi:
10.1108/I
J
P
C
C
-
05
-
2020
-
00
32.
[
2]
G
.
P
r
a
s
a
d
A
c
ha
r
ya
,
M
.
A
s
ha
R
a
ni
,
G
.
G
a
ne
s
h
K
uma
r
,
a
nd
L
.
P
ol
uboyina
,
“
A
da
pt
a
ti
on
of
ma
r
c
h
-
S
S
a
lg
or
it
h
m
to
w
or
d
-
o
r
ie
n
te
d
me
mor
y
bui
lt
-
in
s
e
lf
-
te
s
t
a
nd
r
e
pa
ir
,”
I
ndone
s
ia
n
J
our
nal
of
E
le
c
tr
ic
al
E
ngi
ne
e
r
in
g
and
C
om
put
e
r
Sc
i
e
nc
e
,
vol
.
26,
no.
1,
pp. 96
–
104, Apr
. 2022, doi:
10.11591/i
je
e
c
s
.v26.i1.pp96
-
104.
[
3]
D
.
J
a
ma
l
a
nd
R
.
V
e
e
ti
l,
“
E
f
f
ic
ie
nt
M
B
I
S
T
a
r
e
a
a
nd
te
s
t
ti
me
e
s
ti
ma
to
r
u
s
in
g
ma
c
hi
n
e
le
a
r
ni
ng
t
e
c
hni
que
,”
in
2023
36
th
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on
V
L
SI
D
e
s
ig
n
and
2023
22nd
I
nt
e
r
nat
io
nal
C
onf
e
r
e
n
c
e
on
E
m
be
dde
d
Sy
s
te
m
s
(
V
L
SI
D
)
,
J
a
n.
2023,
pp. 223
–
228
,
doi
:
10.1109/VL
S
I
D
57277.2023.00054.
[
4]
P
.
R
a
ma
kr
is
hna
,
T
.
V
a
ms
hi
ka
,
a
nd
M
.
S
w
a
th
i,
“
F
P
G
A
im
pl
e
me
nt
a
ti
on
of
me
mor
y
B
I
S
T
s
u
s
in
g
s
in
gl
e
in
te
r
f
a
c
e
,”
I
nt
e
r
nat
i
onal
J
our
nal
of
R
e
c
e
nt
T
e
c
hnol
ogy
and E
ngi
ne
e
r
in
g (
I
J
R
T
E
)
, vol
. 9,
no. 3, pp. 55
–
58, S
e
p. 2020, doi:
10.35940/i
jr
te
.B
3975.099320.
[
5]
R
.
S
il
ve
ir
a
,
Q
.
Q
ur
e
s
hi
,
a
nd
R
.
Z
e
li
,
“
F
le
xi
bl
e
a
r
c
hi
te
c
tu
r
e
of
me
mor
y
B
I
S
T
s
,”
in
2018
I
E
E
E
19t
h
L
at
in
-
A
m
e
r
ic
an
T
e
s
t
Sy
m
pos
iu
m
(
L
A
T
S)
, M
a
r
. 2018, pp. 1
–
6
,
doi
:
10.1109/L
A
T
W
.2018.8349666.
[
6]
N
.
A
.
Z
a
ka
r
ia
,
W
.
Z
.
W
.
H
a
s
a
n,
I
.
A
.
H
a
li
n,
R
.
M
.
S
id
e
k,
a
nd
X
.
W
e
n,
“
F
a
ul
t
de
te
c
ti
on
w
it
h
opt
im
um
ma
r
c
h
te
s
t
a
lg
or
it
hm,
”
in
2012
T
hi
r
d
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on
I
nt
e
ll
ig
e
nt
Sy
s
t
e
m
s
M
ode
ll
in
g
and
Si
m
ul
at
io
n
,
F
e
b.
2012,
pp.
700
–
704
,
doi
:
10.1109/I
S
M
S
.2012.88.
[
7]
A
.
Z
.
J
id
in
,
R
.
H
us
s
in
,
W
.
F
.
L
e
e
,
a
nd
M
.
S
.
M
is
pa
n,
“
M
B
I
S
T
i
mpl
e
me
nt
a
ti
on
a
nd
e
va
lu
a
ti
on
in
F
P
G
A
ba
s
e
d
on
lo
w
-
c
ompl
e
xi
ty
ma
r
c
h a
lg
or
it
hms
,”
J
our
nal
of
C
ir
c
ui
ts
, Sy
s
t
e
m
s
and C
o
m
put
e
r
s
, vol
. 33, no. 08, 2023, doi
:
10.1142/S
0218126624501524.
[
8]
I
.
G
.
M
a
tr
i,
A
is
hw
a
r
a
ya
,
N
.
S
hr
e
ya
,
S
.
V
.
S
id
da
ma
l,
a
nd
S
.
V
.
B
udi
ha
l,
“
A
nove
l
ma
r
c
h
X
R
a
lg
or
it
h
m,
de
s
ig
n,
a
nd
te
s
t
a
r
c
hi
te
c
tu
r
e
f
or
me
mo
r
ie
s
,”
in
A
dv
anc
e
s
in
E
le
c
tr
ic
al
an
d
C
om
put
e
r
T
e
c
hnol
ogi
e
s
,
S
pr
in
ge
r
N
a
tu
r
e
S
in
ga
por
e
,
20
22,
pp. 321
–
329
,
doi
:
10.1007/978
-
981
-
19
-
1111
-
8_26.
[
9]
Y
.
X
ia
o,
S
.
L
u,
E
.
W
a
ng,
R
.
Z
hu,
a
nd
Z
.
D
a
i,
“
T
e
s
t
pr
im
it
iv
e
:
a
s
tr
a
ig
ht
f
or
w
a
r
d
me
th
od
to
de
c
oupl
e
ma
r
c
h,”
a
r
X
iv
P
r
e
pr
.
,
pp. 1
–
9, 2023, doi:
10.48550/ar
X
iv
.2309.03214.
[
10]
R
.
W
a
ng,
Z
.
H
ua
ng,
G
.
C
a
i,
a
nd
Z
.
Y
u,
“
A
bui
lt
-
in
s
e
lf
-
te
s
t
c
ir
c
ui
t
ba
s
e
d
on
ma
r
c
h
F
R
D
D
a
lg
or
it
hm
f
or
F
in
F
E
T
me
mo
r
y,”
in
I
ndus
tr
ia
l
E
ngi
ne
e
r
in
g
and
A
ppl
ic
at
io
ns
:
P
r
oc
e
e
di
ngs
of
th
e
10t
h
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on
I
ndus
tr
ia
l
E
ngi
ne
e
r
in
g
and
A
ppl
ic
at
io
ns
(
I
C
I
E
A
2023)
, J
ul
. 2023
,
doi
:
10.3233/AT
D
E
2300
69.
[
11]
V
.
S
.
C
ha
kr
a
va
r
th
i,
A
pr
ac
ti
c
al
appr
oac
h
to
V
L
SI
s
y
s
te
m
on
c
hi
p
(
SoC
)
de
s
ig
n
.
C
ha
m:
S
pr
in
ge
r
I
nt
e
r
na
ti
ona
l
P
ubl
is
hi
ng,
20
22
,
doi
:
10.1007/978
-
3
-
031
-
18363
-
8.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
S
S
N
:
2088
-
8708
I
nt
J
E
lec
&
C
omp
E
ng
,
Vol
.
15
,
No.
1
,
F
e
br
ua
r
y
20
25
:
1
63
-
1
7
3
172
[
12]
T
.
S
.
N
gua
n
K
ong
e
t
al
.
,
“
A
n
e
f
f
ic
ie
nt
M
a
r
c
h
(
5n)
F
S
M
-
ba
s
e
d
me
mor
y
bui
lt
-
in
s
e
lf
-
te
s
t
(
M
B
I
S
T
)
a
r
c
hi
te
c
tu
r
e
,”
in
2021
I
E
E
E
R
e
gi
onal
Sy
m
pos
iu
m
on M
ic
r
o and N
anoe
le
c
tr
oni
c
s
(
R
SM
)
, A
ug. 2021, pp. 76
–
79
,
doi
:
10.1109/R
S
M
52397.2021.9511602.
[
13]
S
. N
. B
a
ge
w
a
di
, S
. S
ha
da
b, a
nd J
. R
oopa
, “
F
a
s
t
B
I
S
T
me
c
ha
ni
s
m f
or
f
a
s
te
r
va
li
da
ti
on of
me
mor
y
a
r
r
a
y,”
i
n
2019 4
th
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on
R
e
c
e
nt
T
r
e
nds
on
E
le
c
t
r
oni
c
s
,
I
nf
or
m
at
io
n,
C
om
m
uni
c
at
io
n
&
T
e
c
hnol
ogy
(
R
T
E
I
C
T
)
,
M
a
y
2019,
pp. 61
–
65
,
doi
:
10.1109/R
T
E
I
C
T
46194.2019.9016882.
[
14]
K
hus
hi
a
nd
K
.
S
in
gh,
“
P
e
r
f
or
ma
nc
e
a
n
a
ly
s
is
of
ma
r
c
h
M
&
B
a
lg
or
it
hms
f
or
me
mor
y
bui
lt
-
in
s
e
lf
-
te
s
t
(
B
I
S
T
)
,”
in
2022
I
E
E
E
W
or
ld
C
onf
e
r
e
nc
e
on A
ppl
ie
d I
nt
e
ll
ig
e
nc
e
and C
om
put
in
g (
A
I
C
)
, 2022, pp. 78
–
84
,
doi
:
10.1109/AI
C
55036.2022.9848869.
[
15]
S
.
B
.
G
ha
le
a
nd
P
.
N
.,
“
D
e
s
ig
n
a
nd
im
pl
e
m
e
nt
a
ti
on
of
me
mo
r
y
B
I
S
T
f
or
hybr
id
c
a
c
he
a
r
c
hi
te
c
tu
r
e
,”
in
2021
6t
h
I
nt
e
r
nat
i
onal
C
onf
e
r
e
nc
e
on C
om
m
uni
c
at
io
n and E
le
c
tr
oni
c
s
Sy
s
te
m
s
(
I
C
C
E
S)
, 2021, pp. 26
–
31
,
doi
:
10.1109/I
C
C
E
S
51350.2021.9489225.
[
16]
J
.
K
r
ut
hi
ka
,
G
.
R
.
N
is
ha
,
R
.
G
a
ya
th
r
i,
a
nd
V
.
J
e
ya
la
ks
hmi
,
“
S
R
A
M
me
mor
y
bui
lt
in
s
e
lf
-
te
s
t
us
in
g
ma
r
c
h
a
lg
or
it
hm
,”
in
2022
I
nt
e
r
nat
io
nal
C
onf
e
r
e
nc
e
on
A
ugm
e
nt
e
d
I
nt
e
ll
ig
e
nc
e
and
Sus
ta
in
abl
e
Sy
s
te
m
s
(
I
C
A
I
SS)
,
N
ov.
2022,
pp.
1288
–
1
292
,
doi
:
10.1109/I
C
A
I
S
S
55157.2022.10010813.
[
17]
I
.
M
r
oz
e
k,
N
.
A
.
S
he
vc
he
nko,
a
nd
V
.
N
.
Y
a
r
mol
ik
,
“
U
ni
v
e
r
s
a
l
a
ddr
e
s
s
s
e
que
nc
e
ge
n
e
r
a
to
r
f
or
me
mor
y
bu
il
t
-
in
s
e
lf
-
t
e
s
t,
”
F
undame
nt
a I
nf
or
m
at
ic
ae
, vol
. 188, no. 1, pp. 41
–
61, De
c
. 202
2, doi:
10.3233/F
I
-
222141.
[
18]
S
.
M
a
r
ti
r
os
ya
n
a
nd
G
.
H
a
r
ut
yunya
n,
“
A
n
e
f
f
ic
ie
nt
f
a
ul
t
de
t
e
c
ti
on
a
nd
di
a
gnos
is
me
th
odol
ogy
f
or
vol
a
ti
le
a
nd
non
-
vol
a
ti
le
me
mor
ie
s
,”
in
2019
C
om
put
e
r
Sc
ie
nc
e
and
I
nf
or
m
at
io
n
T
e
c
hnol
ogi
e
s
(
C
SI
T
)
,
S
e
p.
2019,
pp.
47
–
51
,
doi
:
10.1109/C
S
I
T
e
c
hnol
.2019.8895189.
[
19]
S
.
F
.
A
bd
G
a
ni
,
M
.
F
.
M
is
kon,
R
.
A
.
H
a
mz
a
h,
M
.
S
.
H
a
mi
d,
A
.
F
.
K
a
dmi
n,
a
nd
A
.
I
.
H
e
r
ma
n,
“
S
te
r
e
o
ma
tc
hi
ng
a
lg
or
it
hm
us
in
g
de
e
p
le
a
r
ni
ng
a
nd
e
dge
-
pr
e
s
e
r
vi
ng
f
il
te
r
f
or
ma
c
hi
ne
vi
s
io
n,”
B
ul
le
ti
n
of
E
le
c
tr
ic
al
E
ngi
ne
e
r
in
g
and
I
nf
or
m
at
ic
s
,
vol
.
13,
no.
3,
pp. 1685
–
1693, J
un. 2024, doi:
10.11591/ee
i.
v13i
3.5708.
[
20]
G
.
H
a
r
ut
unya
n,
V
.
A
.
V
a
r
da
ni
a
n,
a
nd
Y
.
Z
o
r
ia
n,
“
M
in
im
a
l
m
a
r
c
h
te
s
ts
f
or
unl
in
ke
d
s
ta
ti
c
f
a
ul
ts
in
r
a
ndom
a
c
c
e
s
s
me
mor
ie
s
,
”
in
23r
d I
E
E
E
V
L
SI
T
e
s
t
Sy
m
pos
iu
m
(
V
T
S’
05)
, 2005, pp. 53
–
59
,
doi
:
10.1109/VT
S
.2005.56.
[
21]
S
.
H
a
mdi
oui
,
“
T
e
s
ti
ng
e
mb
e
dde
d
me
mor
ie
s
:
a
s
ur
ve
y,”
in
M
at
he
m
at
ic
al
and
E
ngi
ne
e
r
in
g
M
e
th
ods
in
C
om
put
e
r
Sc
ie
nc
e
,
A
.
K
uč
e
r
a
,
T
.
A
.
H
e
nz
in
ge
r
,
J
.
N
e
š
e
tř
il
,
T
.
V
oj
na
r
,
a
nd
D
.
A
nt
o
š
,
E
ds
.
B
e
r
li
n:
S
pr
in
ge
r
B
e
r
li
n
H
e
id
e
lb
e
r
g,
2013,
pp.
32
–
42
,
doi
:
10.1007/978
-
3
-
642
-
36046
-
6_4.
[
22]
A
.
Z
.
J
id
in
e
t
al
.
,
“
G
e
ne
r
a
ti
on
of
ne
w
lo
w
-
c
ompl
e
xi
ty
ma
r
c
h
a
lg
or
it
hms
f
or
opt
im
um
f
a
ul
ts
de
te
c
ti
on
in
S
R
A
M
,”
I
E
E
E
T
r
ans
ac
ti
ons
on
C
om
put
e
r
-
A
id
e
d
D
e
s
ig
n
of
I
nt
e
gr
at
e
d
C
ir
c
ui
ts
and
Sy
s
te
m
s
,
vol
.
42,
no.
8,
pp.
2738
–
2751,
20
23,
doi
:
10.1109/T
C
A
D
.2022.3229281.
[
23]
V
.
A
.
V
a
r
da
ni
a
n
a
n
d
Y
.
Z
o
r
i
a
n
,
“
A
M
a
r
c
h
-
ba
s
e
d
f
a
ul
t
l
oc
a
ti
o
n
a
l
go
r
it
h
m
f
o
r
s
t
a
t
ic
r
a
n
do
m
a
c
c
e
s
s
me
mo
r
i
e
s
,
”
in
P
r
oc
e
e
di
n
gs
o
f
th
e
E
ig
ht
h
I
E
E
E
I
n
te
r
na
ti
on
al
O
n
-
L
in
e
T
e
s
ti
n
g
W
or
k
s
h
op
(
I
O
L
T
W
2
00
2
)
,
2
00
2,
pp
.
25
6
–
26
1
,
doi
:
1
0.
11
09
/O
L
T
.2
00
2.
10
30
22
8.
[
24]
X
.
N
in
g,
H
.
Y
a
ng,
M
.
Z
ha
ng,
Y
.
W
a
ng,
Y
.
Z
ha
o,
a
nd
S
.
Q
ia
o,
“
M
ul
ti
-
ty
pe
S
R
A
M
te
s
t
s
tr
uc
tu
r
e
w
it
h
a
n
im
p
r
ove
d
ma
r
c
h
L
R
a
lg
or
it
hm,”
in
2022
I
E
E
E
A
s
ia
P
ac
if
ic
C
onf
e
r
e
nc
e
on
C
ir
c
ui
ts
and
S
y
s
te
m
s
(
A
P
C
C
A
S)
,
N
ov.
2022,
pp.
578
–
582
,
doi
:
10.1109/AP
C
C
A
S
55924.2022.10090328.
[
25]
Z
.
Z
hi
-
c
ha
o,
H
.
L
i
-
ga
ng,
a
nd
W
.
W
u
-
C
he
n
,
“
S
R
A
M
B
I
S
T
d
e
s
ig
n
ba
s
e
d
on
ma
r
c
h
C
+
a
lg
or
it
hm,”
M
od.
E
le
c
tr
on.
T
e
c
h
,
vol
.
34,
no. 10, pp. 149
–
151, 2011.
[
26]
A
.
Z
.
J
id
in
,
R
.
H
us
s
in
,
L
.
W
.
F
ook,
a
nd
M
.
S
.
M
is
pa
n,
“
A
n
a
ut
oma
ti
on
pr
ogr
a
m
f
or
ma
r
c
h
a
lg
or
it
hm
f
a
ul
t
de
te
c
ti
on
a
na
ly
s
is
,
”
i
n
2021
I
E
E
E
A
s
ia
P
ac
if
ic
C
onf
e
r
e
nc
e
on
C
ir
c
ui
t
and
Sy
s
te
m
s
(
A
P
C
C
A
S)
,
2021,
pp.
149
–
152
,
doi
:
10.1109/AP
C
C
A
S
51387.2021.9687806.
B
I
OG
RA
P
HI
E
S
OF
AU
T
HO
RS
A
i
m
a
n
Za
kw
a
n
J
i
d
i
n
recen
t
l
y
co
m
p
l
e
t
ed
h
i
s
P
h
.
D
.
i
n
el
ec
t
ro
n
i
c
e
n
g
i
n
eer
i
n
g
at
U
n
i
v
er
s
i
t
i
Mal
a
y
s
i
a
Perl
i
s
,
Mal
a
y
s
i
a.
H
i
s
res
earc
h
t
o
p
i
c
fo
c
u
s
e
s
o
n
creat
i
n
g
a
n
ew
l
o
w
-
co
mp
l
ex
i
t
y
memo
ry
t
es
t
i
n
g
al
g
o
r
i
t
h
m
fo
r
o
p
t
i
m
u
m
s
t
a
t
i
c
fau
l
t
co
v
er
ag
e
i
n
SRA
M.
Prev
i
o
u
s
l
y
,
h
e
o
b
t
ai
n
ed
h
i
s
M
.
E
n
g
.
in
el
ec
t
ro
n
i
c
a
n
d
mi
cr
o
el
ec
t
ro
n
i
c
s
y
s
t
ems
fr
o
m
E
SIE
E
E
n
g
i
n
eer
i
n
g
Pari
s
,
Fran
ce,
i
n
2
0
1
1
,
b
ef
o
re
w
o
r
k
i
n
g
a
s
a
fi
el
d
-
p
r
o
g
rammab
l
e
g
a
t
e
array
(
FPG
A
)
IP
Co
re
D
es
i
g
n
E
n
g
i
n
eer
a
t
A
l
t
era
Co
r
p
o
ra
t
i
o
n
Ma
l
ay
s
i
a
(
n
o
w
p
art
o
f
In
t
el
).
H
e
i
s
a
fu
l
l
-
t
i
me
l
ec
t
u
rer
an
d
res
earc
h
er
i
n
e
l
ect
r
o
n
i
c
an
d
co
m
p
u
t
er
en
g
i
n
eeri
n
g
a
t
U
n
i
v
er
s
i
t
i
T
e
k
n
i
k
a
l
Mal
ay
s
i
a
Me
l
ak
a
(U
T
eM).
H
i
s
res
earc
h
i
n
t
eres
t
s
i
n
c
l
u
d
e
d
es
i
g
n
fo
r
t
es
t
a
b
i
l
i
t
y
(D
F
T
)
,
v
ery
l
ar
g
e
-
s
cal
e
i
n
t
eg
ra
t
i
o
n
(
V
L
SI
)
,
an
d
f
i
el
d
-
p
r
o
g
ramma
b
l
e
g
at
e
arra
y
(FPG
A
)
s
y
s
t
em
d
e
s
i
g
n
.
H
e
ca
n
b
e
co
n
t
act
e
d
at
emai
l
:
ai
ma
n
za
k
w
a
n
@
u
t
em.
ed
u
.
my
.
R
a
z
a
i
di
Hu
s
s
i
n
recei
v
ed
a
Ph
.
D
.
d
eg
ree
i
n
e
l
ect
r
o
n
i
c
an
d
e
l
ect
r
i
cal
e
n
g
i
n
eer
i
n
g
fro
m
t
h
e
U
n
i
v
ers
i
t
y
o
f
G
l
as
g
o
w
,
t
h
e
U
K
,
i
n
2
0
1
7
w
i
t
h
a
fo
cu
s
o
n
o
x
i
d
e
-
rel
i
ab
i
l
i
t
y
i
s
s
u
e
s
i
n
co
mp
l
emen
t
ary
met
al
-
o
x
i
d
e
-
s
em
i
co
n
d
u
ct
o
r
n
an
o
s
cal
e
d
ev
i
ce
s
.
H
e
j
o
i
n
e
d
U
n
i
v
er
s
i
t
i
Mal
a
y
s
i
a
Perl
i
s
(p
re
v
i
o
u
s
l
y
k
n
o
w
n
a
s
K
U
K
U
M)
i
n
2
0
0
2
.
H
e
i
s
a
fu
l
l
-
t
i
me
as
s
o
c
i
at
e
p
ro
fe
s
s
o
r
at
t
h
e
Facu
l
t
y
o
f
E
l
ect
r
o
n
i
c
E
n
g
i
n
eeri
n
g
T
ec
h
n
o
l
o
g
y
,
U
n
i
v
er
s
i
t
i
Mal
a
y
s
i
a
Perl
i
s
.
H
e
can
b
e
co
n
t
act
e
d
at
emai
l
:
s
h
i
d
ee@
u
n
i
map
.
ed
u
.
my
.
Evaluation Warning : The document was created with Spire.PDF for Python.