I
nte
rna
t
io
na
l J
o
urna
l o
f
E
lect
rica
l a
nd
Co
m
p
ute
r
E
ng
in
ee
ring
(
I
J
E
CE
)
Vo
l.
9
,
No
.
6
,
Dec
em
b
er
201
9
,
p
p
.
5
0
8
4
~5
0
9
4
I
SS
N:
2
0
8
8
-
8708
,
DOI
: 1
0
.
1
1
5
9
1
/
i
j
ec
e
.
v9
i
6
.
p
p
5
0
8
4
-
5094
5084
J
o
ur
na
l ho
m
ep
a
g
e
:
h
ttp
:
//ia
e
s
co
r
e
.
co
m/
jo
u
r
n
a
ls
/in
d
ex
.
p
h
p
/
I
JE
C
E
An ene
rg
y
-
a
w
a
re sy
ste
m
-
on
-
c
hip
ar
chitec
ture
for
i
n
tr
a
prediction
in
H
E
VC s
tanda
rd
E
l A
n
s
a
ri
Abde
s
s
a
m
a
d
1
,
An
a
s
M
a
ns
o
uri
2
,
Ali A
ha
it
o
uf
3
1
,2
F
a
c
u
lt
y
o
f
S
c
ien
c
e
s a
n
d
T
e
c
h
n
o
lo
g
y
,
L
ERS
I,
L
a
b
o
ra
to
ry
,
Un
iv
e
r
sity
o
f
S
id
i
M
o
h
a
m
m
e
d
Be
n
A
b
d
e
ll
a
h
F
e
z
,
M
o
ro
c
c
o
3
Na
ti
o
n
a
l
S
c
h
o
o
l
o
f
A
p
p
li
e
d
S
c
ien
c
e
s,
L
ERS
I,
L
a
b
o
ra
to
ry
,
Un
iv
e
r
sity
o
f
S
id
i
M
o
h
a
m
m
e
d
Be
n
A
b
d
e
ll
a
h
F
e
z
,
M
o
ro
c
c
o
Art
icle
I
nfo
AB
ST
RAC
T
A
r
ticle
his
to
r
y:
R
ec
eiv
ed
Ma
y
1
,
2
0
1
8
R
ev
i
s
ed
J
u
n
2
2
,
2
0
1
9
A
cc
ep
ted
J
u
l 2
,
2
0
1
9
H
ig
h
re
so
lu
ti
o
n
4
K
a
n
d
8
K
a
re
b
e
c
o
m
in
g
th
e
m
o
re
u
se
d
in
v
id
e
o
a
p
p
li
c
a
ti
o
n
s
.
T
h
o
se
re
so
lu
ti
o
n
s
a
re
w
e
ll
su
p
p
o
rted
in
t
h
e
n
e
w
HEV
C
sta
n
d
a
rd
.
T
h
u
s,
e
m
b
e
d
d
e
d
so
l
u
ti
o
n
s
s
u
c
h
a
s
d
e
v
e
lo
p
m
e
n
t
o
f
d
e
d
ica
ted
y
ste
m
s
-
O
n
-
Ch
ip
s
(S
OC)
to
a
c
c
e
lera
te
v
id
e
o
p
ro
c
e
ss
in
g
o
n
o
n
e
c
h
ip
i
n
ste
a
d
o
f
o
n
ly
so
f
t
wa
re
so
lu
ti
o
n
s
a
re
c
o
m
m
e
n
d
a
b
le
.
T
h
is
p
a
p
e
r
p
ro
p
o
s
e
s
a
n
o
v
e
l
p
a
ra
ll
e
l
a
n
d
h
ig
h
e
f
f
ici
e
n
t
h
a
rd
w
a
re
a
c
c
e
lera
to
r
f
o
r
th
e
in
tra
p
re
d
ictio
n
b
lo
c
k
.
T
h
is
a
c
c
e
lera
to
r
a
c
h
iev
e
s
a
h
ig
h
-
sp
e
e
d
trea
tme
n
t
d
u
e
to
p
i
p
e
li
n
e
d
p
ro
c
e
ss
in
g
u
n
it
s
a
n
d
p
a
ra
ll
e
l
sh
a
p
e
d
a
rc
h
it
e
c
tu
re
.
T
h
e
c
o
m
p
lex
it
y
o
f
m
e
m
o
r
y
a
c
c
e
ss
is
a
lso
re
d
u
c
e
d
th
a
n
k
s
to
th
e
p
r
o
p
o
se
d
d
e
sig
n
w
it
h
les
s
in
c
re
a
se
d
p
o
w
e
r
c
o
n
su
m
p
ti
o
n
.
T
h
e
i
m
p
le
m
e
n
tatio
n
w
a
s
p
e
r
f
o
r
m
e
d
o
n
th
e
7
S
e
ries
F
P
G
A
2
8
n
m
tec
h
n
o
lo
g
y
r
e
so
u
rc
e
s
o
n
Zy
n
q
-
7
0
0
0
a
n
d
re
su
lt
s
sh
o
w
,
th
a
t
th
e
p
ro
p
o
se
d
a
rc
h
it
e
c
tu
re
tak
e
s
1
6
5
2
0
L
UT
s
a
n
d
c
a
n
re
a
c
h
1
4
3
.
6
5
M
Hz
a
s
a
m
a
x
i
m
u
m
f
re
q
u
e
n
c
y
a
n
d
it
is
a
b
le
to
su
p
p
o
r
t
th
e
th
ro
u
g
h
p
u
t
o
f
3
8
4
0
×
2
1
6
0
se
q
u
e
n
c
e
a
t
9
0
f
ra
m
e
s
p
e
r
se
c
o
n
d
.
K
ey
w
o
r
d
s
:
FP
GA
Har
d
w
ar
e
HE
VC
I
n
tr
a
p
r
ed
ictio
n
Vid
eo
co
d
in
g
Co
p
y
rig
h
t
©
2
0
1
9
In
stit
u
te o
f
A
d
v
a
n
c
e
d
E
n
g
i
n
e
e
rin
g
a
n
d
S
c
ien
c
e
.
Al
l
rig
h
ts re
se
rv
e
d
.
C
o
r
r
e
s
p
o
nd
ing
A
uth
o
r
:
E
l A
n
s
ar
i
A
b
d
ess
a
m
ad
,
Sid
i M
o
h
a
m
m
ed
B
en
A
b
d
ella
h
Un
iv
er
s
it
y
,
Facu
lt
y
o
f
Sc
ien
ce
s
an
d
T
ec
h
n
o
lo
g
y
o
f
Fez,
L
ab
o
r
ato
r
y
o
f
R
e
n
e
w
ab
le
E
n
e
r
g
y
a
n
d
S
m
ar
t S
y
s
te
m
s
(
L
E
R
S
I
)
,
P
.
O.
B
o
x
2
2
0
2
Fez,
0
0
2
1
2
6
2
6
2
4
3
3
5
5
,
M
o
r
o
cc
o
.
E
m
ail:
ab
d
ess
a
m
ad
.
elan
s
ar
i@
u
s
m
b
a.
ac
.
m
a
1.
I
NT
RO
D
UCT
I
O
N
Hig
h
e
f
f
icie
n
c
y
v
id
eo
co
d
in
g
(
HE
V
C
)
is
t
h
e
la
s
t
v
id
eo
co
d
in
g
s
ta
n
d
ar
d
r
ec
en
tl
y
r
e
leased
an
d
d
ev
elo
p
ed
u
n
d
er
t
h
e
e
f
f
o
r
ts
o
f
J
o
in
t
C
o
llab
o
r
ativ
e
T
ea
m
o
n
Vid
eo
C
o
d
i
n
g
(
J
C
T
-
VC
)
u
n
d
er
t
h
e
I
SO/IE
C
MP
E
G
(
Mo
v
in
g
P
ictu
r
e
E
x
p
er
t
Gr
o
u
p
)
s
tan
d
ar
d
izatio
n
o
r
g
an
izatio
n
a
n
d
th
e
I
T
U
-
T
VC
E
G
(
Vid
eo
C
o
d
in
g
E
x
p
er
t
Gr
o
u
p
)
.
I
t
in
te
g
r
ated
d
if
f
er
en
t
al
g
o
r
ith
m
s
o
f
t
h
e
v
id
eo
co
d
in
g
s
u
ch
a
s
i
n
ter
,
in
tr
a
p
r
ed
ictio
n
,
en
tr
o
p
y
co
d
in
g
an
d
f
ilter
s
.
I
t
ac
h
iev
e
s
g
o
o
d
p
er
f
o
r
m
an
ce
i
m
p
r
o
v
e
m
en
t
w
i
th
an
i
n
cr
ea
s
e
co
m
p
u
tatio
n
al
co
m
p
lex
it
y
co
m
p
ar
ed
w
it
h
its
p
r
ed
ec
ess
o
r
H.
2
6
4
/A
VC
.
Am
o
n
g
t
h
e
i
m
p
o
r
tan
t
f
ea
t
u
r
e
to
o
ls
im
p
le
m
e
n
ted
in
HE
V
C
,
th
e
n
e
w
p
r
o
p
o
s
ed
v
ar
iab
le
s
ize
b
lo
ck
p
r
ed
ictio
n
u
n
it
(
P
U)
u
p
to
6
4
×6
4
d
e
d
icate
d
to
s
u
p
p
o
r
t
v
ar
io
u
s
h
i
g
h
-
r
e
s
o
lu
tio
n
v
id
eo
in
o
p
p
o
s
ite
to
th
e
f
ix
ed
m
ac
r
o
b
lo
ck
1
6
×
16
,
ex
is
tin
g
in
its
p
r
ed
ec
ess
o
r
H.
2
6
4
/A
VC
.
Seco
n
d
l
y
,
th
e
i
n
tr
a
p
r
ed
ictio
n
w
h
ic
h
is
o
n
e
o
f
th
e
i
m
p
o
r
tan
t
an
d
m
o
r
e
ti
m
e
co
n
s
u
m
i
n
g
b
lo
ck
o
f
th
e
HE
V
C
v
id
eo
co
d
in
g
s
tan
d
ar
d
,
u
s
ed
to
co
m
p
u
te
th
e
p
i
x
el
s
p
r
ed
icted
in
s
id
e
th
e
P
Us
.
I
n
th
e
n
e
w
s
ta
n
d
ar
d
,
it
led
s
to
ac
h
i
ev
e
b
etter
ef
f
icie
n
c
y
,
t
h
an
d
ef
i
n
ed
in
th
e
p
r
ev
io
u
s
s
tan
d
ar
d
s
v
id
eo
co
d
in
g
ev
en
if
th
e
s
e
i
m
p
r
o
v
e
m
e
n
ts
m
a
k
e
th
e
al
g
o
r
ith
m
ex
e
cu
tio
n
s
i
g
n
i
f
ican
tl
y
co
m
p
le
x
[
1
,
2]
.
I
n
f
ac
t
,
th
e
in
tr
a
p
r
ed
ictio
n
ex
p
lo
its
th
e
s
p
atial
r
ed
u
n
d
an
c
y
e
x
i
s
ti
n
g
o
n
l
y
in
a
s
in
g
le
f
r
a
m
e
,
it
h
a
s
s
ev
er
al
p
r
ed
ictio
n
m
o
d
e
ls
(
u
p
to
3
5
m
o
d
es)
a
n
d
it
co
m
p
u
te
s
t
h
e
lar
g
e
P
U
b
lo
ck
(
u
p
to
6
4
×6
4
p
ix
els)
.
T
h
is
w
a
y
,
ea
ch
in
tr
a
p
r
ed
ictio
n
b
lo
ck
is
co
m
p
u
ted
b
y
all
p
r
ed
ictio
n
m
o
d
es
a
n
d
th
e
f
i
n
al
ch
o
ice
is
b
ased
o
n
th
e
s
m
alle
s
t
p
r
ed
ictio
n
co
s
t,
co
m
p
u
ted
f
o
r
all
m
o
d
es.
T
h
i
s
p
r
ed
ictio
n
is
p
er
f
o
r
m
ed
f
o
r
b
o
th
C
h
r
o
m
a
a
n
d
L
u
m
a
m
o
d
e
d
ec
is
io
n
.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
A
n
en
erg
y
-
a
w
a
r
e
s
ystem
-
on
-
c
h
ip
a
r
ch
itectu
r
e
fo
r
in
tr
a
p
r
ed
ictio
n
in
HE
V
C
s
ta
n
d
a
r
d
(
A
b
d
ess
a
ma
d
el
A
n
s
a
r
i
)
5085
On
e
o
f
t
h
e
b
i
g
g
e
s
t
b
e
n
ef
i
ts
o
f
u
s
i
n
g
f
ield
p
r
o
g
r
a
m
m
ab
le
g
ate
ar
r
a
y
(
FP
G
A
)
,
to
i
m
p
le
m
en
t
all
o
r
p
ar
tial
p
ar
ts
o
f
t
h
e
HE
V
C
e
n
co
d
er
,
is
to
m
ak
e
av
a
ilab
le
a
ti
m
e
m
ea
s
u
r
i
n
g
u
n
it
in
s
id
e
t
h
e
w
h
o
le
p
r
o
ce
s
s
i
n
g
s
y
s
te
m
u
s
in
g
o
n
l
y
a
s
in
g
le
FP
GA
.
I
n
f
ac
t
t
h
e
co
m
p
le
x
it
y
a
n
al
y
s
e
s
o
f
HE
V
C
h
ar
d
w
ar
e
i
m
p
le
m
en
ta
tio
n
n
ee
d
s
a
r
ig
o
r
o
u
s
ti
m
e
co
n
s
u
m
i
n
g
to
b
e
c
o
n
s
id
er
ed
in
th
e
p
r
o
f
ilin
g
d
u
e
th
e
h
i
g
h
alg
o
r
ith
m
s
co
m
p
lex
i
t
y
o
f
HE
V
C
en
co
d
er
,
an
d
ca
n
b
y
h
elp
f
u
l
f
o
r
th
e
r
ea
l
ti
m
e
p
r
o
ce
s
s
i
n
g
c
h
ec
k
.
On
e
o
f
th
e
la
s
t
g
en
er
atio
n
FP
G
A
f
r
o
m
Xili
n
x
is
Z
y
n
q
-
7000
.
I
t
o
f
f
er
ed
a
s
o
f
t
w
a
r
e
p
r
o
ce
s
s
i
n
g
s
y
s
te
m
o
n
d
u
al
-
co
r
e
A
R
M
co
r
tex
-
A9
an
d
a
h
ar
d
w
ar
e
p
r
o
ce
s
s
in
g
s
y
s
te
m
o
n
P
r
o
g
r
am
m
ab
le
L
o
g
ic
(
P
L
)
in
c
lu
d
i
n
g
on
FP
G
A
i
n
a
s
i
n
g
le
d
ev
ice.
I
t
is
t
h
u
s
p
ar
tic
u
lar
l
y
s
u
itab
le
to
s
t
u
d
y
t
h
e
p
o
s
s
ib
ilit
y
f
o
r
th
e
i
m
p
le
m
en
tatio
n
o
f
t
h
e
HE
VC
w
it
h
i
n
th
is
f
a
m
il
y.
Sin
ce
th
e
f
ir
s
t
s
o
f
t
w
ar
e
v
er
s
i
o
n
o
f
HE
VC
s
ta
n
d
ar
d
w
a
s
p
r
o
p
o
s
ed
,
m
an
y
s
o
f
t
w
ar
e
an
d
h
ar
d
w
ar
e
i
m
p
le
m
en
ta
tio
n
s
h
a
v
e
b
ee
n
d
ev
elo
p
ed
to
o
f
f
er
s
tati
s
tical
an
al
y
s
es
co
n
ce
r
n
in
g
th
e
e
x
e
cu
tio
n
ti
m
e,
p
o
w
er
co
n
s
u
m
p
tio
n
an
d
co
m
p
le
x
it
y
an
al
y
s
es i
n
o
r
d
er
to
r
ea
ch
r
ea
l
ti
m
e
p
r
o
ce
s
s
i
n
g
f
o
r
th
e
h
i
g
h
er
r
eso
lu
tio
n
s
li
k
e
4
K
an
d
8
K.
I
n
T
h
is
w
o
r
k
,
w
e
f
o
cu
s
ed
o
n
a
h
ar
d
w
ar
e
i
m
p
le
m
en
tatio
n
o
f
t
h
e
I
n
tr
a
p
r
ed
ictio
n
b
lo
ck
o
f
t
h
e
HE
V
C
en
co
d
er
o
n
th
e
Fiel
d
-
P
r
o
g
r
a
m
m
ab
le
Gate
A
r
r
a
y
(
FP
G
A
)
.
T
h
e
in
tr
a
p
r
ed
ictio
n
al
g
o
r
it
h
m
co
m
p
u
te
s
all
p
r
ed
icted
p
ix
els
u
s
i
n
g
s
e
v
er
al
m
o
d
es.
T
h
e
in
tr
a
p
r
ed
ictio
n
alg
o
r
ith
m
is
i
n
s
er
te
d
in
b
o
th
J
M
an
d
HM
r
ef
er
en
c
e
s
o
f
t
w
ar
e
en
co
d
er
f
o
r
th
e
s
ta
n
d
ar
d
s
H.
2
6
4
/A
VC
an
d
HE
VC
r
esp
ec
ti
v
el
y
.
So
f
t
w
ar
e
p
r
o
f
ili
n
g
of
all
f
u
n
ctio
n
s
u
s
ed
f
o
r
th
e
i
n
tr
a
m
o
d
e
p
r
ed
ictio
n
t
h
at
allo
w
s
f
o
r
a
co
m
p
ar
is
o
n
b
et
w
ee
n
t
h
e
co
m
p
lex
i
t
ies
o
f
al
l
f
u
n
ctio
n
s
in
t
h
e
w
h
o
le
e
n
co
d
er
HE
VC
is
p
r
esen
ted
i
n
th
e
w
o
r
k
.
Mo
s
t
o
f
p
o
r
tab
le
d
ev
ices
s
u
p
p
o
r
ted
ap
p
licatio
n
s
th
at
ar
e
d
ed
icate
d
f
o
r
m
u
lti
m
ed
ia
lik
e
s
tr
ea
m
i
n
g
f
r
o
m
o
n
l
in
e
s
er
v
ice
s
.
L
o
t
o
f
t
h
ese
d
ev
ice
s
ar
e
h
eter
o
g
e
n
eo
u
s
e
m
b
ed
d
ed
s
y
s
te
m
s
an
d
ar
e
d
esig
n
ed
in
s
p
ir
it
o
f
lo
w
p
o
w
er
co
n
s
u
m
p
tio
n
.
T
h
e
Z
y
n
q
S
y
s
te
m
O
n
C
h
ip
(
S
OC
)
,
f
o
r
m
Xili
n
x
,
is
o
n
e
o
f
th
e
s
e
d
ev
ice
s
t
h
at
co
m
b
i
n
e
h
eter
o
g
e
n
eo
u
s
e
m
b
ed
d
ed
s
y
s
te
m
i
n
cl
u
d
in
g
i
n
t
h
e
s
a
m
e
ti
m
e
a
P
r
o
g
r
a
m
m
a
b
le
L
o
g
ic
(
P
L
)
an
d
P
r
o
ce
s
s
in
g
S
y
s
te
m
(
P
S)
o
n
o
n
e
C
h
ip
w
it
h
v
ar
io
u
s
s
y
s
te
m
o
f
co
m
m
u
n
icatio
n
s
w
it
h
th
e
m
,
c
o
n
s
eq
u
e
n
tl
y
,
lo
t
o
f
r
ep
o
r
ted
w
o
r
k
in
t
h
e
liter
at
u
r
e
g
iv
e
t
h
e
co
m
p
u
tatio
n
al
co
m
p
lex
it
y
i
n
t
h
is
C
h
ip
.
T
h
e
au
th
o
r
s
P
en
g
et
al
.
[
3
]
p
r
o
p
o
s
ed
an
d
i
m
p
le
m
e
n
t
ed
a
h
i
g
h
l
y
in
teg
r
ated
ac
ce
ler
ati
n
g
s
o
l
u
tio
n
f
o
r
-
b
o
d
y
M
O
ND
s
i
m
u
latio
n
s
o
n
v
ar
io
u
s
p
r
o
ce
s
s
o
r
s
lik
e
AR
M
an
d
GP
U,
an
d
o
n
th
e
FP
GA
Z
y
n
q
-
7020
So
C
in
teg
r
ated
u
s
in
g
b
o
th
s
i
m
p
lifie
d
p
ip
elin
e
an
d
b
an
d
w
id
th
r
eq
u
ir
e
m
en
t
tech
n
iq
u
e
s
.
T
h
eir
p
r
o
p
o
s
ed
s
o
lu
tio
n
o
n
t
h
e
SO
C
h
a
s
a
g
o
o
d
p
e
r
f
o
r
m
a
n
ce
in
p
o
w
er
co
n
s
u
m
p
tio
n
u
p
to
1
0
ti
m
es a
n
d
th
e
b
est p
er
f
o
r
m
a
n
ce
in
ter
m
s
o
f
co
s
t o
v
er
5
0
%
[
4
,
5
]
.
R
e
m
a
in
d
er
o
f
t
h
is
p
ap
er
is
o
r
g
an
ized
as f
o
llo
w
s
.
Sectio
n
I
I
g
iv
es a
r
elate
d
w
o
r
k
o
f
i
n
tr
a
p
r
ed
ictio
n
in
HE
VC
,
s
ec
tio
n
I
I
I
d
escr
ib
es
th
e
in
tr
a
p
r
ed
ictio
n
alg
o
r
ith
m
i
n
v
id
eo
en
co
d
i
n
g
HE
V
C
s
tan
d
ar
d
.
Sectio
n
I
V
is
d
ed
icate
d
to
th
e
p
r
o
p
o
s
ed
h
ar
d
w
ar
e
ar
ch
itect
u
r
e
f
o
r
ev
er
y
m
o
d
e.
T
h
e
s
y
n
t
h
es
is
a
n
d
p
er
f
o
r
m
an
ce
r
es
u
lts
o
b
tain
ed
in
th
e
FP
G
A
a
n
d
AS
I
C
w
i
th
co
m
p
ar
is
o
n
to
o
t
h
er
w
o
r
k
s
a
n
d
p
r
esen
ted
i
n
s
ec
tio
n
V.
Fi
n
all
y
,
th
e
co
n
cl
u
s
io
n
an
d
f
u
t
u
r
e
w
o
k
s
ar
e
g
iv
e
n
i
n
Sectio
n
VI
.
2.
RE
L
AT
E
D
WO
RK
Ma
n
y
r
esear
c
h
w
o
k
s
h
a
v
e
b
ee
n
p
u
b
lis
h
ed
in
t
h
e
liter
at
u
r
e
in
o
r
d
er
to
p
r
o
p
o
s
e
an
ef
f
icie
n
t
h
ar
d
w
ar
e
ar
ch
itect
u
r
e
o
n
l
y
f
o
r
i
n
tr
a
p
r
ed
ictio
n
in
th
e
p
r
ev
io
u
s
a
n
d
la
s
t
HE
VC
s
tan
d
ar
d
[6
–
14]
,
h
o
w
ev
er
,
a
s
t
u
d
y
o
f
th
is
alg
o
r
ith
m
o
n
s
y
s
te
m
o
n
C
h
ip
h
as
n
o
t
y
et
b
ee
n
r
ep
o
r
ted
in
o
u
r
k
n
o
w
led
g
e.
Fo
r
h
ar
d
w
ar
e
i
m
p
le
m
en
ta
tio
n
s
s
ev
er
al
w
o
r
k
s
h
a
v
e
p
r
o
p
o
s
ed
an
ac
ce
ler
atin
g
d
esi
g
n
f
o
r
in
tr
a
p
r
ed
ictio
n
,
i
m
p
le
m
en
t
ed
o
n
FP
GA
-
b
ased
tech
n
o
lo
g
y
i
n
[6
-
11]
,
o
r
o
n
an
A
SIC
c
h
ip
s
in
[
1
2
]
,
o
r
in
b
o
th
A
SI
C
an
d
FP
G
A
in
[
1
3
,
1
4
]
.
So
m
e
o
f
th
e
s
e
d
esig
n
s
ar
e
p
r
o
p
o
s
ed
to
s
u
p
p
o
r
t a
ll
m
o
d
e
d
ec
is
io
n
an
d
all
P
U
b
lo
ck
s
izes.
I
n
[
6
]
,
th
e
h
ar
d
w
ar
e
i
m
p
le
m
e
n
tatio
n
o
f
HE
V
C
d
ec
o
d
er
in
clu
d
in
g
t
h
e
i
n
tr
a
p
r
ed
ictio
n
o
n
FP
GA
i
s
p
r
esen
ted
,
it
ca
n
b
e
u
s
ed
f
o
r
d
ec
o
d
in
g
i
n
r
ea
l
ti
m
e
4
K
r
eso
l
u
tio
n
w
it
h
3
0
f
r
a
m
e
p
er
s
ec
o
n
d
(
f
p
s
)
.
T
h
e
au
th
o
r
s
u
s
ed
o
p
ti
m
izatio
n
av
ailab
le
o
n
FP
GA
o
n
t
h
e
p
ip
elin
e
s
ta
g
es,
th
er
ef
o
r
e,
th
e
y
u
s
e
a
s
in
g
le
c
y
cle
r
ef
er
en
ce
p
ix
el
p
r
o
ce
s
s
in
g
i
n
t
h
e
in
tr
a
p
r
ed
ictio
n
a
n
d
an
o
t
h
er
tec
h
n
iq
u
es
f
o
r
b
lo
ck
s
tr
a
n
s
f
o
r
m
an
d
b
o
t
h
f
ilter
in
s
id
e
HE
V
C
d
ec
o
d
e,
s
o
th
at
lead
to
a
h
ig
h
t
h
r
o
u
g
h
p
u
t.
in
[
7
]
,
a
m
u
l
tip
le
tech
n
iq
u
e
s
ar
e
p
r
o
p
o
s
ed
s
o
as
to
m
a
k
e
h
ar
d
w
ar
e
ac
ce
ler
ato
r
d
ev
elo
p
e
d
f
o
r
in
tr
a
p
r
ed
ictio
n
o
f
HE
V
C
f
u
n
ctio
n
i
n
g
in
f
u
ll
p
ip
elin
e.
T
h
e
t
h
r
ee
t
ec
h
n
iq
u
es
ad
o
p
ted
ar
e
n
o
v
el
b
u
f
f
er
s
tr
u
ct
u
r
e
f
o
r
r
ef
er
en
ce
s
a
m
p
le
s
,
m
o
d
e
d
ep
en
d
en
t
s
ca
n
n
in
g
o
r
d
er
an
d
an
in
v
er
s
e
m
e
th
o
d
f
o
r
r
ef
er
en
ce
s
a
m
p
les
ex
te
n
s
io
n
,
th
at
a
llo
w
ed
to
p
r
o
ce
s
s
4
p
ix
els
p
er
co
ck
c
y
c
le,
t
h
er
ef
o
r
e,
th
e
t
h
r
o
u
g
h
p
u
t
o
f
t
h
is
ar
ch
it
ec
tu
r
e
ca
n
s
u
p
p
o
r
t
Qu
ad
Fu
ll HD
(
3
8
4
0
×
2
1
6
0
)
at
3
0
f
r
am
e
s
/s
.
a
n
d
also
th
e
ac
ce
ler
ato
r
ca
n
co
m
p
u
te
all
i
n
tr
a
p
r
ed
ictio
n
m
o
d
es.
I
n
[
8
]
.
a
h
ar
d
w
ar
e
f
o
r
in
tr
a
p
r
ed
ictio
n
is
p
r
esen
ted
f
o
r
p
r
o
ce
s
s
in
g
all
m
o
d
es,
it
is
b
as
ed
o
n
t
w
o
tech
n
iq
u
es
to
r
ed
u
ce
th
e
co
m
p
u
tatio
n
co
m
p
lex
it
y
a
n
d
to
ac
ce
ler
ate
tr
ea
t
m
en
t.
T
h
e
f
ir
s
t
is
ad
o
p
ted
f
o
r
an
g
u
lar
m
o
d
e
s
an
d
it
ca
l
led
P
r
o
ce
s
s
in
g
E
le
m
e
n
t
f
o
r
An
g
u
lar
(
P
E
A
)
m
o
d
e
s
an
d
t
h
e
s
ec
o
n
d
i
s
ca
lled
P
r
o
ce
s
s
in
g
ele
m
e
n
t
f
o
r
p
lan
ar
(
P
E
R
)
m
o
d
e,
th
en
th
e
d
esi
g
n
is
s
tr
u
c
tu
r
ed
in
r
ep
ea
ted
p
ath
s
i
n
o
r
d
er
to
co
m
p
u
te
i
n
p
ar
allel
w
a
y
.
T
h
er
ef
o
r
e,
th
e
ar
ch
itect
u
r
e
ca
n
p
r
o
ce
s
s
1
0
8
0
p
@
1
0
0
o
r
4
K@
2
4
in
r
ea
l ti
m
e.
I
n
[
9
]
,
th
e
h
ar
d
w
ar
e
ac
ce
ler
at
o
r
s
f
o
r
ea
c
h
s
ize
b
lo
ck
f
r
o
m
4
×4
to
3
2
×3
2
an
d
a
d
esi
g
n
ar
e
p
r
esen
ted
w
it
h
t
h
ese
s
y
n
t
h
esi
s
r
es
u
lts
.
T
h
e
h
ar
d
w
ar
e
is
p
r
o
p
o
s
ed
o
n
VHDL
la
n
g
u
a
g
e
a
n
d
it
i
m
p
le
m
en
ted
o
n
th
e
Xili
n
x
Vir
tex
-
6
th
at
is
m
a
n
u
f
ac
tu
r
e
d
at
th
e
4
0
n
m
tech
n
o
lo
g
y
n
o
d
e.
T
h
e
d
esig
n
tak
e
s
1
7
0
K
L
UT
s
an
d
1
1
0
K
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
9
,
No
.
6
,
Dec
em
b
er
2
0
1
9
:
5
0
8
4
-
509
4
5086
r
eg
is
ter
s
w
it
h
o
u
t
a
n
y
b
lo
ck
o
f
m
e
m
o
r
y
f
r
o
m
a
v
ailab
le
r
e
s
o
u
r
ce
s
.
Af
ter
T
h
e
s
y
n
t
h
e
s
is
d
esig
n
,
t
h
e
m
a
x
i
m
u
m
f
r
e
q
u
en
c
y
u
p
to
2
1
9
MH
z
an
d
is
ca
p
ab
le
to
co
m
p
u
te
2
4
f
p
s
f
o
r
4
K
r
eso
lu
tio
n
.
J
ian
g
et
al.
[
1
0
]
d
ev
elo
p
ed
th
eir
w
o
r
k
in
a
p
ip
elin
ed
w
a
y
a
n
d
i
m
p
le
m
en
ted
it
o
n
a
Xilin
x
Vir
tex
5
d
ev
ice.
I
t
co
n
s
u
m
ed
6
9
k
L
U
T
s
,
6
4
in
ter
n
al
D
SP
s
an
d
1
4
8
k
b
it
i
n
ter
n
al
m
e
m
o
r
ies
t
y
p
e
B
R
A
M.
Kah
n
e
t
al.
in
[
1
1
]
p
r
o
p
o
s
ed
a
p
ip
elin
ed
d
esig
n
t
h
at
co
n
tain
th
e
p
ar
allel
p
r
o
ce
s
s
in
g
E
le
m
e
n
t
s
(
P
E
)
s
u
p
p
o
r
tin
g
all
P
U
s
izes
an
d
m
o
d
es
d
ec
is
io
n
,
it
ac
h
ie
v
es
2
1
3
MH
z
a
n
d
ca
n
w
o
r
k
in
r
ea
l
ti
m
e
co
n
s
tr
ain
t
s
ac
h
ie
v
i
n
g
1
2
0
f
p
s
f
o
r
1
0
8
0
r
eso
lu
tio
n
a
n
d
3
0
f
p
s
f
o
r
th
e
4
k
r
eso
lu
tio
n
.
I
n
p
ap
er
[
1
2
]
,
a
VL
SI
ar
ch
ite
ctu
r
e
is
p
r
esen
ted
f
o
r
in
tr
a
p
r
ed
ictio
n
in
HE
VC
s
ta
n
d
ar
d
.
T
h
e
d
esig
n
ad
o
p
ted
th
r
ee
tech
n
iq
u
es
to
r
ed
u
ce
s
y
s
te
m
atica
ll
y
t
h
e
co
m
p
lex
i
t
y
.
T
h
e
f
ir
s
t
is
t
h
e
in
t
eg
r
ated
h
ier
ar
ch
ical
m
e
m
o
r
y
i
n
s
tead
o
f
g
ates
as
r
eg
i
s
ter
s
u
s
ed
f
o
r
s
to
r
i
n
g
th
e
n
ei
g
h
b
o
r
in
g
s
a
m
p
le
s
t
h
at
ca
n
i
n
cr
ea
s
e
th
e
th
r
o
u
g
h
p
u
t.
T
h
e
s
ec
o
n
d
is
a
m
o
d
e
-
ad
ap
tiv
e
s
c
h
ed
u
li
n
g
s
c
h
e
m
e
t
h
at
lead
to
p
r
o
v
id
e
at
least
2
s
a
m
p
les
/c
y
cle
s
.
T
h
e
t
h
r
ee
is
r
ed
u
cin
g
t
h
e
m
u
ltip
lier
s
b
y
s
h
a
r
in
g
th
e
m
in
th
e
p
r
o
p
o
s
ed
ar
c
h
itect
u
r
e.
Usi
n
g
t
h
i
s
tech
n
iq
u
e
ca
n
r
ed
u
ce
t
h
e
co
n
s
u
m
p
t
io
n
ar
ea
,
b
u
t,
th
ese
t
h
r
ee
tech
n
iq
u
e
s
co
n
s
u
m
e
m
o
r
e
p
o
w
er
d
u
e
to
e
m
p
lo
y
ed
S
R
A
M
m
e
m
o
r
y
a
n
d
th
e
p
o
s
t
la
y
o
u
t
s
i
m
u
latio
n
s
h
o
w
s
a
p
o
w
er
co
n
s
u
m
p
ti
o
n
u
p
to
2
.
1
1
m
W
.
T
h
e
d
esig
n
is
s
y
n
t
h
esized
at
2
0
0
MH
z
u
s
in
g
A
SI
C
tech
n
o
lo
g
y
4
0
n
m
p
r
o
ce
s
s
an
d
ca
n
s
u
p
p
o
r
ted
th
e
r
eso
lu
tio
n
3
8
4
0
×
2
1
6
0
@
3
0
in
r
ea
l ti
m
e
p
r
o
ce
s
s
in
g
.
I
n
[
1
3
]
,
b
o
th
tech
n
iq
u
es
p
ix
e
l
eq
u
alit
y
a
n
d
p
ix
el
s
i
m
ilar
it
y
a
r
e
u
s
ed
in
o
r
d
er
to
r
e
d
u
ce
th
e
a
m
o
u
n
t
o
f
th
e
p
er
f
o
r
m
ed
ca
lcu
latio
n
b
y
t
h
e
in
tr
a
p
r
ed
ictio
n
al
g
o
r
i
th
m
i
n
HE
VC
.
C
o
n
s
eq
u
en
tl
y
,
t
h
e
e
n
er
g
y
co
n
s
u
m
p
tio
n
is
also
r
ed
u
ce
d
.
T
h
e
h
ar
d
w
a
r
e
d
esig
n
s
p
er
f
o
r
m
ed
o
n
l
y
4
×4
an
d
8
×8
an
g
u
lar
p
r
ed
ictio
n
m
o
d
e
s
a
n
d
i
t
i
m
p
le
m
en
ted
o
n
b
o
th
Xili
n
x
Vir
tex
6
FP
G
A
w
it
h
1
5
0
MH
z
o
f
f
r
eq
u
en
c
y
,
a
n
d
o
n
A
SI
C
tech
n
o
lo
g
y
u
s
i
n
g
9
0
n
m
s
tan
d
ar
d
ce
ll
lib
r
ar
y
f
r
o
m
S
y
n
o
p
s
y
s
,
th
er
e
f
o
r
e,
th
e
y
ca
n
p
r
o
ce
s
s
3
0
f
r
am
e
p
er
s
ec
o
n
d
f
o
r
f
u
ll
HD
(
1
9
2
0
×1
0
8
0
)
r
eso
lu
tio
n
f
o
r
t
w
o
ca
s
es.
I
n
[
1
4
]
,
a
s
y
m
m
e
tr
ic
p
r
o
p
r
iet
y
o
f
i
n
tr
a
p
er
d
itio
n
eq
u
atio
n
s
o
f
th
e
h
o
r
izo
n
tal
an
d
v
er
tical
d
ir
ec
tio
n
s
is
e
x
p
lo
ited
,
th
at
lead
to
l
i
m
it
e
s
t
u
d
y
o
n
e
d
ir
ec
tio
n
an
d
it
ch
an
g
e
th
e
o
r
d
er
o
f
t
h
e
s
e
lecte
d
r
ef
er
en
ce
s
s
a
m
p
les
f
o
r
th
e
o
th
er
d
ir
ec
tio
n
.
T
h
e
h
a
r
d
w
ar
e
d
esig
n
is
i
m
p
le
m
e
n
ted
o
n
b
o
th
Vir
tex
6
FP
G
A
f
r
o
m
Xilin
x
t
h
at
w
o
r
k
s
at
2
3
4
MH
z
an
d
T
SMC
1
8
0
n
m
C
MO
S
p
r
o
ce
s
s
tec
h
n
o
l
o
g
y
w
i
th
t
h
e
f
r
eq
u
e
n
c
y
o
p
er
atio
n
is
2
1
8
MH
Z
.
T
h
e
au
th
o
r
s
p
r
o
p
o
s
e
an
ar
ch
it
ec
tu
r
e
f
o
r
4
×4
,
8
×8
,
1
6
×
1
6
a
n
d
3
2
×
3
2
an
g
u
lar
p
r
ed
ictio
n
m
o
d
e
s
o
n
l
y
o
n
th
e
HE
VC
s
ta
n
d
ar
d
,
th
eir
p
r
o
p
o
s
ed
ar
ch
itectu
r
e
is
b
ased
o
n
D
SP
b
lo
ck
in
s
id
e
i
n
FP
G
A
(
Xi
lin
x
XC
6
V
L
X7
5
T
FF
1
7
5
9
FP
GA
)
.
th
eir
o
p
ti
m
izatio
n
,
lead
to
3
4
.
6
6
%
less
en
er
g
y
co
n
s
u
m
p
tio
n
th
a
n
t
h
e
o
r
ig
i
n
al
FP
G
A
i
m
p
le
m
en
ta
tio
n
(
i.e
.
w
i
th
o
u
t
i
m
p
le
m
e
n
tatio
n
o
n
t
h
e
D
S
P
b
lo
ck
)
o
f
HE
VC
in
tr
a
p
r
ed
ictio
n
an
d
t
h
is
ar
ch
itect
u
r
e,
h
ar
d
w
ar
e
ca
n
b
e
p
r
o
ce
s
s
ed
5
5
f
u
ll h
d
1
0
8
0
p
(
1
9
2
0
×1
0
8
0
)
f
r
am
e
p
er
s
ec
o
n
d
.
3.
H
E
VC
I
N
T
RA
P
R
E
DI
CT
I
O
N
AL
G
O
RI
T
H
M
3
.
1
.
O
v
er
v
ie
w
o
f
H
E
VC
e
nco
der
A
d
escr
ip
tio
n
o
f
t
h
e
w
h
o
le
H
E
VC
en
co
d
er
b
l
o
ck
d
iag
r
a
m
i
s
g
i
v
e
n
i
n
F
ig
u
r
e
1
.
T
h
e
m
a
i
n
m
o
d
u
les
ar
e
th
e
s
a
m
e
in
to
p
r
ev
io
u
s
s
t
an
d
ar
d
s
,
b
u
t
HE
VC
co
m
es
with
f
le
x
ib
ili
t
y
co
d
in
g
b
lo
ck
ag
ain
s
t
e
x
is
t
in
g
f
i
x
ed
an
d
les
s
co
m
p
le
x
it
y
m
o
d
u
le
s
p
r
o
p
o
s
ed
f
o
r
HE
VC
e
n
co
d
er
is
k
n
o
w
n
b
y
h
i
g
h
co
m
p
le
x
it
y
t
h
a
n
d
ec
lar
ed
s
tan
d
ar
d
s
.
Ge
n
er
all
y
,
it
r
ec
eiv
es
a
n
i
n
p
u
t
YUV
f
r
a
m
es
t
y
p
e
s
(
o
r
GOP
[
1
5
]
)
an
d
g
e
n
er
ate
s
a
f
ter
tr
ea
t
m
en
t
a
b
its
tr
ea
m
d
ata
in
its
o
u
tp
u
t
s
id
es.
A
b
r
ief
d
e
s
cr
ip
tio
n
o
f
t
h
ese
m
o
d
u
le
s
i
n
s
id
e
HE
V
C
e
n
co
d
er
is
g
i
v
i
n
g
i
n
t
h
is
w
o
r
k
,
an
d
f
o
r
m
o
r
e
d
etails
r
ea
d
er
ca
n
r
etu
r
n
,
f
o
r
in
s
ta
n
ce
,
to
r
ef
.
[
2
]
.
T
r
a
n
s
f
o
r
m
&
Q
u
a
n
t
i
z
a
t
i
o
n
E
n
t
r
o
p
y
C
o
d
i
n
g
I
n
v
e
r
s
e
&
Q
u
a
n
t
i
z
a
t
i
o
n
I
n
t
r
a
p
r
e
d
i
c
t
i
o
n
I
n
t
e
r
p
r
e
d
i
c
t
i
o
n
D
e
b
l
o
k
i
n
g
F
i
l
t
e
r
F
r
a
m
e
s
B
u
f
f
e
r
S
a
m
p
l
e
A
d
a
p
t
i
v
e
O
f
f
s
e
t
(
S
A
O
)
-
+
+
+
I
n
p
u
t
f
r
a
m
e
b
i
t
s
t
r
e
a
m
Fig
u
r
e
1
.
B
lo
ck
-
d
iag
r
a
m
o
f
t
h
e
HE
VC
E
n
co
d
er
w
i
th
t
h
e
m
a
i
n
m
o
d
u
le
s
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
A
n
en
erg
y
-
a
w
a
r
e
s
ystem
-
on
-
c
h
ip
a
r
ch
itectu
r
e
fo
r
in
tr
a
p
r
ed
ictio
n
in
HE
V
C
s
ta
n
d
a
r
d
(
A
b
d
ess
a
ma
d
el
A
n
s
a
r
i
)
5087
T
h
e
s
ev
er
al
b
lo
c
k
s
ar
e:
1
)
I
n
t
er
p
r
ed
ictio
n
:
th
e
I
n
ter
p
r
ed
ictio
n
i
s
u
s
ed
to
r
ed
u
ce
e
x
ch
a
n
g
e
b
et
w
ee
n
s
u
cc
e
s
s
i
v
e
p
ict
u
r
es
i
n
th
e
Y
UV
s
eq
u
e
n
ce
b
y
d
eter
m
in
at
i
o
n
o
f
th
e
p
r
ed
icted
b
lo
ck
f
o
r
th
e
c
u
r
r
en
t
p
ictu
r
e
av
ailab
le
in
t
h
e
d
ec
o
d
ed
p
ictu
r
e
b
u
f
f
er
.
I
t
is
p
r
o
c
ess
i
n
g
i
n
f
i
r
s
t
b
y
m
o
tio
n
co
m
p
e
n
s
a
ted
p
r
ed
ictio
n
,
b
ased
o
n
m
o
tio
n
v
ec
to
r
s
an
d
it
f
o
llo
w
e
d
b
y
s
u
b
-
s
a
m
p
le
I
n
ter
p
o
latio
n
f
ilter
i
n
g
.
2
)
T
h
e
T
r
an
s
f
o
r
m
a
n
d
in
v
er
s
e
tr
an
s
f
o
r
m
ar
e
r
ef
er
en
ce
d
to
T
U
to
d
eter
m
in
e
th
e
s
ize
o
f
m
atr
ix
r
a
n
g
i
n
g
f
r
o
m
4
×
4
to
3
2
×
3
2
to
b
e
u
s
e
d
to
tr
ea
t
th
e
r
esid
u
al
m
atr
ices
a
f
ter
p
r
ed
ictio
n
p
r
o
ce
s
s
.
3
)
T
h
e
q
u
an
tizatio
n
a
n
d
in
v
er
s
e
q
u
an
t
izati
o
n
ar
e
also
s
i
m
ilar
to
ex
is
ti
n
g
in
th
e
o
ld
s
ta
n
d
ar
d
H.
2
6
4
/A
VC
.
T
h
is
o
p
er
at
io
n
s
ca
les
u
p
to
en
co
d
e
co
ef
f
icien
ts
j
u
s
t
a
f
ter
tr
an
s
f
o
r
m
u
s
i
n
g
b
y
e
n
tr
o
p
y
c
o
d
i
n
g
an
d
to
d
ec
o
d
e
th
e
m
f
o
r
f
ilter
b
lo
ck
.
4
)
E
n
tr
o
p
y
E
n
co
d
in
g
:
T
h
e
E
n
tr
o
p
y
en
co
d
in
g
u
s
ed
to
d
ec
o
d
e
th
e
in
p
u
t b
it
s
tr
ea
m
C
o
n
te
x
t
A
d
ap
ti
v
e
B
in
ar
y
A
r
ith
m
et
ic
C
o
d
in
g
(
C
A
B
AC
)
,
w
h
ich
i
s
th
e
o
n
l
y
u
s
ed
tec
h
n
iq
u
e
f
o
r
th
i
s
d
ec
o
d
in
g
p
r
o
ce
d
u
r
e,
t
h
at
is
s
a
m
e
a
s
f
o
n
d
ed
i
n
th
e
h
i
g
h
p
r
o
f
ile
i
n
to
H.
2
6
4
/A
VC
[
1
6
]
.
5
)
Deb
lo
c
k
in
g
f
ilter
&
Sa
m
p
le
-
ad
ap
tiv
e
Of
f
s
e
t:
T
h
e
Deb
lo
ck
in
g
Fil
ter
an
d
Sa
m
p
le
-
A
d
ap
tiv
e
O
f
f
s
e
t
ar
e
t
w
o
f
i
lter
s
d
esig
n
ed
to
i
m
p
r
o
v
e
co
d
in
g
ef
f
ic
ien
c
y
af
ter
th
e
HE
VC
p
r
o
ce
s
s
o
r
s
an
d
th
e
y
r
e
m
o
v
e
t
h
e
ed
g
e
e
s
s
e
n
tiall
y
a
f
f
ec
ted
b
y
d
ec
is
io
n
m
o
d
es
o
n
o
n
e
p
ictu
r
e,
w
h
ich
i
s
s
to
r
ed
i
n
a
b
u
f
f
er
f
o
r
u
s
e
it
in
th
e
ca
lc
u
latio
n
o
f
t
h
e
m
o
ti
o
n
esti
m
a
tio
n
.
6
)
T
h
e
in
tr
a
p
r
ed
ictio
n
m
o
d
es
in
HE
VC
ar
e
f
i
x
ed
in
3
5
m
o
d
es,
w
h
i
c
h
ar
e
DC
,
p
lan
ar
an
d
3
3
an
g
u
lar
m
o
d
es.
T
h
e
DC
a
n
d
p
lan
ar
ar
e
n
a
m
ed
m
o
d
e
0
an
d
1
r
esp
ec
tiv
el
y
.
T
h
e
p
r
ed
ictio
n
m
o
d
e
2
to
1
8
ar
e
th
e
f
ir
s
t
is
t
h
e
h
al
f
-
p
a
r
t
y
h
o
r
izo
n
ta
l
an
g
u
lar
p
r
ed
ictio
n
m
o
d
es,
a
n
d
th
e
s
ec
o
n
d
is
v
er
tical
m
o
d
es
f
r
o
m
1
9
to
3
4
.
T
h
e
in
tr
a
p
r
ed
ictio
n
is
ap
p
lied
to
m
o
d
e
L
u
m
a
an
d
C
h
r
o
m
a
b
lo
ck
s
,
w
h
ic
h
ar
e
s
ized
f
r
o
m
3
2
×3
2
d
o
w
n
to
4
×4
p
ix
els.
T
h
er
ef
o
r
e,
th
e
eq
u
atio
n
s
p
er
f
o
r
m
th
e
HE
VC
i
n
tr
a
p
r
ed
ictio
n
alg
o
r
ith
m
th
at
is
i
m
p
le
m
e
n
te
d
in
H
M
(
HE
V
C
te
s
t
m
o
d
el)
o
f
HE
V
C
r
e
f
er
en
ce
e
n
co
d
er
i
s
g
iv
e
n
b
elo
w
w
it
h
ea
ch
m
o
d
e
in
[
1
7
]
.
3
.
2
.
Co
din
g
s
t
ruct
ure
C
u
r
r
en
tl
y
,
th
e
HE
VC
i
s
w
o
r
k
in
g
in
h
i
g
h
l
y
f
lex
ib
le
a
n
d
e
f
f
i
cien
t
b
lo
ck
co
d
i
n
g
p
o
r
tio
n
i
n
g
s
tr
u
ct
u
r
e
w
h
ic
h
is
d
i
v
id
ed
in
to
f
o
u
r
lev
els,
g
i
v
i
n
g
a
s
f
o
llo
w
s
:
C
T
U,
C
U,
P
U,
an
d
T
U.
First,
ea
ch
f
r
a
m
e
is
d
iv
id
ed
in
to
b
lo
ck
s
n
a
m
ed
co
d
in
g
tr
ee
u
n
it
(
C
T
Us)
th
at
h
a
v
e
v
ar
io
u
s
s
ize
s
i
n
to
1
6
×1
6
,
3
2
×
3
2
o
r
6
4
×
6
4
.
T
h
e
C
T
U
in
cl
u
d
es
o
n
e
co
d
in
g
tr
ee
b
lo
ck
(
C
T
B
)
f
o
r
L
u
m
a
a
n
d
t
w
o
C
T
B
s
f
o
r
co
m
p
o
n
e
n
t
C
h
r
o
m
a.
T
h
e
C
T
U
is
s
i
m
ilar
to
f
i
x
ed
Ma
cr
o
b
lo
ck
1
6
×1
6
ex
is
t
in
g
i
n
t
h
e
p
r
ec
ed
en
t
s
ta
n
d
ar
d
H.
2
6
4
/A
VC
.
T
h
en
,
th
e
co
d
in
g
u
n
it
(
C
U)
is
u
n
d
er
C
T
U
an
d
it
ca
n
b
e
s
p
lit
in
to
f
o
u
r
d
ep
t
h
s
a
s
s
h
o
w
n
i
n
Fig
u
r
e
2
,
it
i
s
s
ized
a
s
q
u
ar
e
r
eg
io
n
8
×8
,
1
6
×1
6
,
3
2
×3
2
an
d
6
4
×
6
4
p
ix
els
d
ep
e
n
d
in
g
o
n
r
e
s
o
lu
tio
n
o
f
s
eq
u
e
n
ce
s
co
d
in
g
.
Ne
x
t,
T
h
e
C
U
its
e
lf
ca
n
b
e
d
iv
id
ed
i
n
P
r
ed
ictio
n
Un
it
(
P
U)
.
T
h
is
b
lo
ck
ca
n
b
e
u
s
ed
f
o
r
t
h
e
i
n
tr
a
an
d
in
ter
b
lo
ck
p
r
ed
ictio
n
w
it
h
s
ize
s
r
an
g
i
n
g
f
r
o
m
4
×
4
d
o
w
n
to
6
4
×6
4
ac
co
r
d
in
g
to
m
o
d
e
d
ec
is
io
n
.
Fi
n
all
y
,
th
e
tr
an
s
f
o
r
m
U
n
it
(
T
U)
is
d
eter
m
i
n
ed
f
r
o
m
P
U
an
d
it su
i
tab
le
f
o
r
tr
an
s
f
o
r
m
a
n
d
q
u
an
t
if
icatio
n
.
Fig
u
r
e
2
.
E
x
a
m
p
le
o
f
C
U
in
c
o
d
in
g
tr
ee
s
tr
u
c
tu
r
e
(
C
T
U)
3
.
3
.
DC
m
o
de
s
T
h
e
DC
m
o
d
es
is
d
ed
icate
d
to
im
a
g
es
t
h
at
h
a
v
e
a
litt
le
ch
a
n
g
e
b
et
w
ee
n
th
e
n
ei
g
h
b
o
r
in
g
p
ix
els,
th
i
s
m
o
d
e
is
m
o
r
e
co
n
v
e
n
ien
t
to
g
en
er
ate
a
n
at
u
r
al
p
r
ed
ictio
n
b
lo
ck
[
1
]
.
T
h
e
DC
m
o
d
e
is
b
ased
o
n
all
n
eig
h
b
o
r
in
g
v
er
tical
a
n
d
h
o
r
izo
n
tal
ad
j
ac
en
t
p
i
x
els.
T
h
e
p
r
ed
ictio
n
is
ca
lcu
lated
u
s
i
n
g
eq
.
(
1
)
f
o
r
th
e
D
C
m
o
d
e
o
f
p
ix
el
(
x
,
y
)
an
d
eq
.
(
2
)
f
o
r
th
e
t
h
r
ee
p
ix
el
s
(
0
,
0
)
,
(
x
,
0
)
an
d
(
0
,
y
)
.
T
h
o
s
e
eq
u
atio
n
s
ar
e
g
iv
e
n
a
s
f
o
llo
w
s
:
p
p
r
e
d
(
x
,
y
)
is
th
e
v
alu
e
o
f
th
e
p
r
ed
ictio
n
s
a
m
p
le
p
ix
e
l
w
it
h
x
,
y
=
0
…
Nc
-
1
p
r
ef
(
x
,
y
)
is
th
e
v
alu
e
o
f
th
e
n
ei
g
h
b
o
u
r
i
n
g
s
a
m
p
le
p
i
x
el
w
it
h
x
,
y
=0
…
Nc
∗
2
−
1
dc
Va
l
=
(
∑
p
r
ef
(
x
,
−
1
)
N
c
−
1
x
=
0
+
∑
p
r
ef
(
−
1
,
y
)
N
c
−
1
y
=
0
)
≫
(
l
og2
(
Nc
)
+
1
)
p
p
r
ed
(
x
,
y
)
=
dc
Va
l
(
1
)
8
×
8
3
2
x
3
2
16
×
16
16
×
16
4
×
4
4
×
4
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
9
,
No
.
6
,
Dec
em
b
er
2
0
1
9
:
5
0
8
4
-
509
4
5088
W
ith
Nc
=n
T
b
s
p
r
esen
ted
s
ize
o
f
th
e
tr
an
s
f
o
r
m
b
lo
ck
.
p
p
r
ed
(
0
,
0
)
=
(
p
r
ef
(
−
1
,
0
)
+
2
∗
dc
Va
l
+
p
r
ef
(
0
,
−
1
)
+
2
)
≫
2
p
p
r
e
d
(
x
,
0
)
=
(
p
r
ef
(
x
,
−
1
)
+
3
∗
dc
Va
l
+
2
)
≫
2
p
p
r
ed
(
0
,
y
)
=
(
p
r
ef
(
−
1
,
y
)
+
3
∗
dc
Va
l
+
2
)
≫
2
(
2
)
3.
4
.
P
l
a
na
r
m
o
de
T
h
e
HE
VC
ad
o
p
ted
also
o
t
h
er
m
o
d
e
n
a
m
ed
P
lan
ar
in
o
r
d
er
to
s
o
lv
e
th
e
p
r
o
b
lem
ca
u
s
ed
b
y
th
e
an
g
u
lar
p
r
ed
ictio
n
m
o
d
es
th
at
h
a
v
e
a
b
ad
r
esu
lts
w
h
e
n
th
e
P
Us
ar
e
s
a
m
p
le
s
p
ath
n
o
t
r
elate
d
to
an
y
d
ir
ec
tio
n
.
T
h
e
p
lan
ar
m
o
d
e
is
co
m
p
u
ted
b
ased
o
n
f
o
u
r
s
a
m
p
l
e
r
ef
er
en
ce
p
ix
el
s
an
d
g
i
v
i
n
g
b
y
eq
u
atio
n
(
3
)
:
p
p
r
ed
(
x
,
y
)
=
(
(
Nc
−
1
−
x
)
∗
p
r
ef
(
−
1
,
y
)
∗
p
r
ef
(
Nc
,
−
1
)
+
(
Nc
−
1
−
y
)
∗
p
r
ef
(
x
,
−
1
)
+
(
y
+
1
)
∗
p
r
ef
(
−
1
,
Nc
)
+
Nc
)
≫
(
l
og2
(
Nc
)
+
1
)
(
3
)
W
h
e
re
p
re
d
(x
,
y
)
is t
h
e
p
re
d
icte
d
s
a
m
p
le co
rre
sp
o
n
d
in
g
t
o
l
o
c
a
ti
o
n
(
x
,
y
)
a
n
d
Nc
is t
h
e
p
re
d
icti
o
n
u
n
i
t
siz
e
a
n
d
th
e
v
a
lu
e
o
f
x
a
n
d
y
a
re
f
ro
m
0
to
Nc
-
1.
3.
5
.
T
he
a
ng
ula
r
m
o
des
T
h
e
an
g
u
lar
p
r
ed
ictio
n
m
o
d
es
ar
e
p
r
o
v
id
ed
f
o
r
co
m
p
lex
te
x
t
u
r
e
an
d
h
i
g
h
f
r
eq
u
e
n
c
y
co
m
p
o
n
en
t,
it
i
s
ap
p
lied
to
3
3
d
if
f
er
e
n
t
d
ir
ec
ti
o
n
s
t
h
at
u
s
ed
t
h
e
n
e
ig
h
b
o
r
in
g
p
ix
els,
t
h
ese
d
ir
ec
tio
n
s
ar
e
h
o
r
izo
n
tal
(
2
to
1
8
)
in
cl
u
d
in
g
h
o
r
i
zo
n
ta
l
m
o
d
e
(
1
0
)
an
d
v
er
tical
(
1
9
to
3
4
)
also
i
n
clu
d
i
n
g
th
e
v
er
tical
m
o
d
e
(
2
6
)
,
T
h
ese
d
ir
ec
tio
n
s
ar
e
g
iv
e
n
i
n
F
ig
u
r
e
3
.
E
ac
h
s
a
m
p
le
is
p
r
ed
icted
u
s
i
n
g
p
r
o
j
ec
tio
n
o
f
i
ts
p
o
s
itio
n
o
n
th
e
s
et
r
ef
er
en
ce
s
a
m
p
le
s
ar
r
ay
,
s
o
t
h
e
p
r
ed
ictio
n
s
a
m
p
l
es is
ca
lcu
lated
b
y
th
e
f
o
llo
w
i
n
g
eq
u
atio
n
s
:
iIdx
=
(
(
x
+
1
)
∗
Int
r
a
Pr
e
dAngl
e
)
≫
5
(
5
)
iF
a
c
t
=
(
(
x
+
1
)
∗
Int
r
a
Pr
e
d
A
n
gl
e
)
&
31
(
6
)
p
e
rd
Sampl
e
[
x
]
[
y
]
=
(
(
32
−
iFa
ct
)
∗
re
f
[
y
+
iId
x
+
1
]
+
iFa
ct
∗
re
f
[
y
+
iId
x
+
2
]
+
16
)
≫
5
(
7
)
W
h
er
e
p
r
ed
s
am
p
le
i
s
th
e
p
r
ed
icted
p
ix
el
attac
h
ed
to
p
o
s
itio
n
(
x
,
y
)
.
An
d
in
tr
ap
r
ed
An
g
le
is
r
elate
d
to
th
e
in
tr
a
p
r
ed
ictio
n
m
o
d
e
o
r
d
ir
ec
tio
n
,
r
ef
[
+
+
2
]
c
o
r
r
esp
o
n
d
s
to
a
r
ef
er
en
ce
s
i
m
p
le
ar
r
ay
,
th
i
s
eq
u
atio
n
p
r
o
ce
s
s
th
e
v
er
tical
m
o
d
e
s
,
to
co
m
p
u
te
th
e
h
o
r
izo
n
tal
m
o
d
es,
w
e
ch
a
n
g
e
th
e
p
o
s
itio
n
o
f
x
b
y
y
a
n
d
v
ice
v
er
s
a
.
Fig
u
r
e
3
.
T
h
e
d
if
f
er
en
t d
ir
ec
ti
o
n
f
o
r
3
3
an
g
u
lar
m
o
d
es d
ir
ec
tio
n
0
-
5
-
10
-
15
-
20
-
25
-
30
-
30
-
25
-
20
-
15
-
10
-
5
0
5
10
15
20
25
30
5
10
15
20
25
30
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
A
n
en
erg
y
-
a
w
a
r
e
s
ystem
-
on
-
c
h
ip
a
r
ch
itectu
r
e
fo
r
in
tr
a
p
r
ed
ictio
n
in
HE
V
C
s
ta
n
d
a
r
d
(
A
b
d
ess
a
ma
d
el
A
n
s
a
r
i
)
5089
4.
P
RO
P
O
SE
D
ARCH
I
T
E
CT
URE
T
h
e
p
r
o
p
o
s
ed
d
esig
n
s
u
p
p
o
r
ts
all
s
izes
b
lo
c
k
d
ef
i
n
ed
f
o
r
i
n
tr
a
p
r
ed
ictio
n
in
HE
V
C
s
ta
n
d
a
r
d
,
a
s
4
×
4
to
3
2
×
3
2
co
m
i
n
g
f
r
o
m
t
h
e
s
iz
e
r
an
g
e
o
f
P
U
t
h
at
ar
e
6
4
×
6
4
d
o
w
n
to
4
×
4
.
T
h
e
o
v
er
all
v
ie
w
o
f
t
h
e
p
r
o
p
o
s
ed
ar
ch
itect
u
r
e
is
p
r
esen
ted
i
n
Fig
u
r
e
4
.
I
t
i
s
co
m
p
o
s
ed
o
f
4
m
aj
o
r
b
lo
ck
s
w
it
h
a
co
n
tr
o
l
u
n
it
.
T
h
e
f
ir
s
t
is
th
e
b
lo
ck
d
ed
icate
f
o
r
m
e
m
o
r
y
,
t
h
is
ar
ch
itec
tu
r
e
lead
s
to
lo
ad
in
g
an
d
s
to
r
ag
e
o
f
t
h
e
n
ei
g
h
b
o
r
in
g
p
ix
el
f
r
o
m
o
n
e
p
ar
t
,
an
d
to
s
to
r
ag
e
th
e
p
ix
el
co
m
i
n
g
f
r
o
m
t
h
e
u
n
i
t p
r
ed
ictio
n
p
r
o
ce
s
s
f
o
r
f
r
o
m
t
h
e
o
th
er
p
ar
t
.
S
h
o
w
n
i
n
F
ig
u
r
e
5
,
t
h
e
s
ec
o
n
d
b
lo
ck
id
is
th
e
D
C
i
n
tr
a
p
r
ed
ictio
n
m
o
d
e.
I
t
co
n
s
i
s
ts
o
f
a
d
d
er
s
an
d
s
h
i
f
ter
to
p
er
f
o
r
m
ca
lc
u
lat
io
n
o
f
(
1
)
an
d
(
2
)
.
T
h
en
an
y
r
ed
u
ctio
n
o
f
t
h
e
co
m
p
u
tatio
n
a
l
c
o
m
p
le
x
it
y
co
m
in
g
w
it
h
t
h
is
ar
c
h
itect
u
r
e
ca
n
b
e
ex
p
lo
ited
f
o
r
th
e
h
u
g
e
s
ize
b
lo
ck
tr
ea
t
m
e
n
t.
Fig
u
r
e
4
.
T
h
e
g
lo
b
al
ar
ch
itectu
r
e
f
o
r
co
m
p
u
ti
n
g
th
e
i
n
tr
a
p
r
ed
ictio
n
.
A
l
g
o
r
it
h
m
i
n
HE
V
C
en
co
d
er
Fig
u
r
e
5
.
T
h
e
p
r
o
p
o
s
ed
a
r
ch
itecu
tr
e
f
o
r
I
n
tr
a
p
r
ed
ictio
n
DC
p
r
o
ce
s
s
in
g
w
i
th
6
4
in
p
u
t sa
m
p
les
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
9
,
No
.
6
,
Dec
em
b
er
2
0
1
9
:
5
0
8
4
-
509
4
5090
T
h
e
t
w
o
la
s
t
b
lo
ck
s
co
n
ce
r
n
p
lan
ar
an
d
a
n
g
u
lar
m
o
d
e
p
r
ed
icti
o
n
.
Fo
r
th
e
s
e
t
w
o
m
o
d
es,
w
e
u
s
e
th
e
p
r
o
ce
s
s
i
n
g
u
n
i
t
p
r
esen
ted
in
Fi
g
u
r
e
6
.
T
h
is
d
esi
g
n
p
er
f
o
r
m
s
ca
lc
u
latio
n
d
escr
ib
ed
b
y
(
3
)
to
(
7
)
an
d
it
co
n
s
is
ts
o
f
m
u
ltip
lex
er
,
ad
d
er
s
an
d
t
w
o
P
E
.
T
h
e
in
p
u
t
d
ata
ar
e
r
ec
eiv
ed
f
r
o
m
t
h
e
n
ei
g
h
b
o
r
in
g
p
ix
el
s
a
n
d
r
esu
lt
s
S
1
an
d
S
2
ar
e
s
u
p
p
lied
r
esp
ec
tiv
el
y
f
o
r
t
h
e
p
lan
ar
an
d
an
g
u
lar
m
o
d
e
.
T
h
e
co
n
tr
o
l
u
n
it
s
h
o
w
n
i
n
Fi
g
u
r
e
4
is
ad
d
ed
in
o
r
d
er
to
m
an
a
g
e
t
h
e
g
lo
b
al
s
y
s
t
e
m
o
f
th
e
w
h
o
le
d
esig
n
.
I
t
in
clu
d
es
t
h
e
o
r
g
an
i
za
tio
n
o
f
s
a
m
p
le
s
in
p
u
ts
an
d
o
u
tp
u
ts
f
o
r
ea
ch
ca
s
e
o
f
s
izes
P
Us
b
lo
ck
s
.
also
it
ap
p
lied
in
lo
o
p
af
ter
clo
c
k
c
y
cles
co
m
p
u
ted
b
y
a
co
u
n
ter
.
I
t
a
ls
o
u
s
ed
to
co
n
tr
o
l
t
h
e
d
esig
n
n
o
t
s
h
o
w
n
i
n
F
ig
u
r
e
4
to
s
elec
t th
e
b
est
m
o
d
e
co
m
p
ar
is
o
n
b
et
w
ee
n
all
p
r
ed
icted
s
a
m
p
les.
P
E
P
1
P
E
+
+
+
P
2
P
3
P
4
P
5
P
6
P
7
P
8
A
1
A
2
A
3
A
4
n
t
1
6
1
6
>
>
5
L
o
g
2
(
n
t
+
1
)
>
>
S
1
S
2
Fig
u
r
e
6
.
T
h
e
p
r
o
p
o
s
ed
h
ar
d
w
ar
e
ac
ce
ler
ato
r
p
r
o
ce
s
s
in
g
ele
m
en
t (
P
E
)
f
o
r
co
m
p
u
ti
n
g
t
h
e
b
o
th
I
n
tr
a
p
r
ed
ictio
n
p
lan
ar
an
d
an
g
u
lar
As
s
h
o
w
n
i
n
Fi
g
u
r
e
7
,
th
e
r
ef
er
en
ce
p
ix
e
ls
v
alu
e
s
m
an
a
g
e
m
e
n
t
f
o
r
in
tr
a
p
r
ed
ictio
n
p
r
o
ce
s
s
in
g
,
all
r
ef
er
en
ce
s
a
m
p
les
ar
e
f
r
o
m
t
h
e
ab
o
v
e
a
n
d
le
f
t
o
f
t
h
e
e
x
is
t
in
g
P
U,
t
h
at
ar
e
s
to
r
ed
in
m
e
m
o
r
y
B
lo
ck
R
A
M
(
B
R
A
M)
in
s
tead
o
f
r
eg
i
s
ter
s
,
w
h
ic
h
ca
n
r
ed
u
ce
t
h
e
co
m
p
le
x
it
y
a
s
s
o
ciate
d
to
ac
ce
s
s
a
d
at
a,
ev
e
n
i
f
th
e
p
o
w
er
h
as
litt
le
in
cr
ea
s
ed
.
T
h
e
r
e
f
er
en
ce
s
a
m
p
les
ar
e
o
r
o
r
g
an
ize
d
in
s
u
ch
a
w
a
y
t
h
at
ar
e
a
v
a
ilab
le
in
o
n
e
ar
r
a
y
,
w
it
h
t
h
e
f
ir
s
t
av
ailab
le
r
ef
er
e
n
ce
s
a
m
p
le
at
t
h
e
lef
t
ar
e
r
ep
r
esen
ted
b
y
t
h
e
r
ef
er
en
ce
s
a
m
p
le
v
er
tical
a
n
d
th
e
s
e
co
n
d
p
ar
t a
v
ailab
le
r
ig
h
t
b
y
t
h
e
h
o
r
izo
n
ta
l r
ef
er
en
ce
s
a
m
p
les.
Fig
u
r
e
7
.
P
lace
m
e
n
t
t
h
e
n
ei
g
h
b
o
u
r
in
g
p
i
x
els i
n
th
e
R
A
M
m
e
m
o
r
y
.
P
x
in
d
icate
s
th
e
p
i
x
el
with
x
=
0
,
.
.
n
T
b
S *
2
−
1
an
d
n
T
b
S is
eq
u
al
to
th
e
len
g
t
h
ti
m
e
s
th
e
w
id
t
h
o
f
t
h
e
b
lo
ck
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
A
n
en
erg
y
-
a
w
a
r
e
s
ystem
-
on
-
c
h
ip
a
r
ch
itectu
r
e
fo
r
in
tr
a
p
r
ed
ictio
n
in
HE
V
C
s
ta
n
d
a
r
d
(
A
b
d
ess
a
ma
d
el
A
n
s
a
r
i
)
5091
T
h
e
P
r
o
ce
s
s
in
g
P
ar
t
(
P
P)
in
F
ig
u
r
e
5
is
a
3
2
u
n
s
ig
n
ed
b
it
s
m
u
ltip
lier
b
ased
o
n
t
h
e
C
ar
r
y
S
av
e
A
d
d
er
p
r
o
p
o
s
ed
b
y
S
in
g
h
et
al.
i
n
wo
r
k
[
1
8
]
.
T
h
en
w
e
av
o
id
u
s
i
n
g
DSP
s
k
n
o
w
n
b
y
th
eir
h
i
g
h
p
o
w
er
co
n
s
u
m
p
t
io
n
,
ev
en
w
i
th
s
o
m
e
i
n
cr
ea
s
e
o
f
t
h
e
co
n
s
u
m
ed
ar
ea
.
T
h
e
g
lo
b
al
d
esig
n
i
n
F
ig
u
r
e
4
is
d
esig
n
ed
i
n
a
w
a
y
to
h
av
e
a
h
ig
h
er
t
h
r
o
u
g
h
p
u
t
b
y
ex
p
lo
itin
g
th
e
av
ailab
le
p
ip
elin
e
an
d
e
m
p
lo
y
ed
d
ee
p
p
ar
allelis
m
to
ac
h
iev
e
r
ea
l
ti
m
e
p
r
o
ce
s
s
in
g
.
R
eg
a
r
d
in
g
o
n
l
y
t
h
e
r
ate
r
eq
u
ir
e
m
en
ts
,
it
ca
n
b
e
an
t
ici
p
ated
th
at
b
o
th
p
ip
elin
e
an
d
d
ee
p
er
p
ar
allelis
m
i
n
cl
u
d
ed
in
th
e
in
tr
a
p
r
ed
ictio
n
d
esig
n
attain
t
h
e
h
i
g
h
est
r
ate.
Nev
er
th
ele
s
s
,
th
e
s
tag
e
o
r
p
ar
allelis
m
i
s
o
n
l
y
b
e
n
ef
icial
i
f
m
an
y
lo
g
ic
o
u
tp
u
t
p
ath
s
in
h
ar
d
w
ar
e
ar
c
h
i
tectu
r
e
,
an
d
th
e
n
e
w
in
p
u
ts
s
a
m
p
le
s
ar
e
av
ailab
le
af
ter
s
ev
er
al
clo
ck
c
y
cle
s
,
h
o
w
ev
er
,
th
e
p
o
w
er
co
n
s
u
m
p
tio
n
i
s
lo
w
er
w
h
e
n
o
n
e
co
n
s
id
er
t
h
e
p
o
w
er
r
eq
u
ir
e
m
e
n
t
fo
r
t
h
e
p
ar
allel
ar
ch
itect
u
r
e
in
s
tead
o
f
p
ip
elin
e
th
a
t
n
ee
d
s
m
o
r
e
r
eso
u
r
ce
s
ar
ea
.
B
ec
au
s
e
o
f
th
o
s
e
ab
o
v
e
m
e
n
tio
n
ed
p
r
o
p
er
ties
,
it
n
ec
ess
ar
y
to
in
clu
d
e
t
h
e
p
ip
elin
e
an
d
p
ar
allelis
m
s
ta
g
e
o
n
w
h
o
le
i
n
tr
a
p
r
ed
ictio
n
d
esig
n
.
T
h
e
ad
v
an
tag
e
o
f
t
h
e
p
ip
elin
e
is
m
o
r
e
d
etailed
i
n
w
o
r
k
[
1
9
]
,
w
h
ic
h
t
h
e
au
t
h
o
r
s
ar
e
u
s
ed
a
f
ast
VL
SI
p
ip
elin
ed
m
u
l
tip
ler
less
f
ix
ed
-
p
o
in
t
8
x
8
DC
T
.
T
h
ese
ar
ch
itect
u
r
es c
an
b
e
im
p
r
o
v
ed
w
h
en
ap
p
lied
2
-
s
tag
e
p
ip
elin
e
t
h
an
n
o
n
-
p
ip
elin
e.
T
h
e
s
o
lu
tio
n
u
s
ed
i
n
t
h
is
w
o
r
k
is
a
s
f
o
llo
w
s
:
f
ir
s
t
s
tag
e
i
s
ex
p
lo
ited
f
o
u
r
P
E
s,
th
at
i
s
p
r
o
ce
s
s
ed
in
th
e
p
ar
allelis
m
(
h
a
s
f
o
u
r
i
n
p
u
t
s
a
m
p
les
f
o
r
ea
ch
ca
s
e
o
f
P
E
in
th
e
s
a
m
e
ti
m
e)
f
o
r
th
e
p
u
r
p
o
s
e
o
f
p
r
ed
icted
f
o
u
r
o
u
tp
u
t
s
a
m
p
le
s
v
a
lu
e
s
,
th
en
p
ip
elin
e
s
ta
g
e
is
ap
p
lied
to
co
m
p
u
te
all
r
e
m
a
in
i
n
g
p
r
ed
icted
o
u
tp
u
ts
s
a
m
p
les
o
n
ea
ch
s
izes
P
Us.
O
u
r
h
ar
d
war
e
ar
ch
itectu
r
e
w
o
r
k
s
in
p
ar
a
llel
w
it
h
f
o
u
r
le
v
el
s
a
m
p
le
s
a
n
d
tak
e
s
o
n
l
y
e
ig
h
t
s
a
m
p
le
clo
ck
c
y
cle,
i
n
clu
d
i
n
g
th
e
co
n
s
u
m
ed
ti
m
e
w
it
h
l
o
ad
in
g
an
d
s
to
r
in
g
d
ata
b
et
w
ee
n
th
e
ex
ter
n
al
m
e
m
o
r
ie
s
in
o
r
d
er
to
h
av
e
r
ea
l ti
m
e
p
r
o
ce
s
s
i
n
g
.
T
h
e
p
r
o
p
o
s
ed
p
ar
allel
p
r
o
ce
s
s
in
g
f
o
r
t
h
e
s
m
a
ll 4
×4
in
tr
a
p
r
ed
ictio
n
b
lo
ck
as
s
h
o
w
n
i
n
F
ig
u
r
e
5
an
d
6
is
ex
p
lo
ited
in
a
p
ip
elin
ed
p
r
o
ce
d
u
r
e
to
p
r
o
ce
s
s
b
ig
g
er
b
lo
ck
8
×8
,
1
6
×1
6
an
d
3
2
×
3
2
.
T
h
is
allo
w
s
an
in
cr
ea
s
e
o
f
t
h
e
b
it
s
tr
ea
m
w
it
h
a
d
r
a
wb
ac
k
o
f
s
o
m
e
in
cr
ea
s
e
in
h
ar
d
w
ar
e
r
eso
u
r
ce
s
b
y
lo
ad
in
g
a
n
d
p
r
o
ce
s
s
in
g
d
ata
at
th
e
s
a
m
e
ti
m
e.
T
h
e
m
o
d
e
D
C
p
r
ed
icted
all
s
a
m
p
les
b
ase
d
o
n
th
e
d
esi
g
n
i
n
F
ig
u
r
e
5
,
w
h
ic
h
h
a
v
e
as
i
n
p
u
t
th
e
r
ef
er
e
n
ce
s
a
m
p
le
n
ei
g
h
b
o
r
in
g
a
v
ailab
le
i
n
B
R
A
M
p
[
i]
[
0
]
an
d
p
[
0
]
[
j
]
w
it
h
x
i
a
n
d
y
j
,
r
esp
ec
tiv
el
y
.
I
t p
r
o
p
o
s
ed
o
n
tr
ee
to
c
o
m
p
u
te
3
2
x
3
2
s
ize
b
lo
ck
an
d
its
s
m
a
l
l o
n
e.
5.
E
XP
E
R
I
M
E
NT
A
L
RE
SUL
T
S
T
h
e
p
r
o
p
o
s
ed
d
esig
n
b
lo
ck
f
o
r
th
e
in
tr
a
p
r
ed
ictio
n
in
HE
V
C
is
f
ir
s
tl
y
co
d
ed
in
a
h
i
g
h
-
le
v
el
lan
g
u
a
g
e
C
++
in
o
r
d
er
to
v
alid
ate
an
d
r
ed
u
ce
th
e
co
m
p
u
ta
tio
n
al
co
m
p
lex
i
t
y
.
Seco
n
d
l
y
,
t
h
is
d
esi
g
n
i
s
i
m
p
le
m
e
n
ted
in
VHDL
lan
g
u
ag
e,
th
e
n
t
h
e
h
ar
d
w
ar
e
ar
ch
itec
tu
r
e
p
r
o
to
t
y
p
ed
o
n
A
r
tex
-
7
FP
G
A
,
t
h
at
i
s
i
n
t
eg
r
ated
o
n
Z
YNQ
-
7
0
0
0
p
r
o
ce
s
s
o
r
th
at
is
x
c7
z0
2
0
-
1
clg
4
0
0
f
r
o
m
Xili
n
x
an
d
ac
ce
ler
atio
n
h
ar
d
w
ar
e
w
as
v
e
r
if
ied
o
n
Mic
r
o
ze
d
.
T
ab
le
1
s
h
o
w
s
t
h
e
s
y
n
th
e
s
i
s
r
esu
lt
s
o
f
t
h
e
g
lo
b
al
ar
ch
itec
tu
r
e
p
r
o
p
o
s
ed
in
Fig
u
r
e
4
.
T
h
e
m
ax
i
m
u
m
f
r
eq
u
e
n
c
y
attain
ed
i
s
1
4
3
MH
z.
On
e
b
lo
ck
R
AM
w
a
s
u
s
ed
f
o
r
s
to
r
ag
e
th
e
n
ei
g
h
b
o
r
in
g
p
ix
e
ls
a
n
d
als
o
f
o
r
th
e
s
to
r
ag
e
o
f
all
p
o
s
s
ib
le
o
u
tp
u
t
i
n
tr
a
p
r
ed
ictio
n
p
ix
el
s
r
es
u
lts
.
T
h
e
p
ix
els
g
e
n
er
ated
af
ter
t
h
e
p
r
o
p
o
s
ed
h
ar
d
w
ar
e
i
n
tr
a
p
r
ed
ictio
n
ar
ch
itectu
r
e
p
r
o
ce
s
s
in
g
ar
e
f
ir
s
t
co
m
p
ar
ed
to
p
r
ed
icted
p
ix
els
co
m
i
n
g
f
r
o
m
t
h
e
r
ef
er
en
ce
s
o
f
t
w
ar
e
HE
VC
HM
.
1
5
.
0
v
er
s
io
n
.
T
h
e
n
th
e
y
al
s
o
co
m
p
ar
ed
w
ith
s
o
f
t
w
ar
e
i
m
p
le
m
e
n
ted
o
n
lan
g
u
ag
e
C
++
to
v
alid
ate
th
e
p
r
o
p
o
s
ed
ar
ch
itectu
r
e.
A
ll
th
ese
co
m
p
ar
i
s
o
n
s
s
h
o
w
a
g
o
o
d
ar
g
u
m
e
n
t.
5
.
1
.
Reso
urce
s
a
na
ly
s
is
T
h
e
p
r
o
p
o
s
ed
ar
ch
itectu
r
e
w
a
s
s
y
n
t
h
esized
u
s
i
n
g
r
ec
e
n
t
d
esig
n
en
v
ir
o
n
m
e
n
t
(
Xili
n
x
Vi
v
ad
o
2
0
1
6
.
3
)
.
T
h
e
r
esu
lt
s
ar
e
p
r
es
en
ted
i
n
T
ab
le
1
.
T
h
is
ar
ch
it
ec
tu
r
e
w
o
r
k
s
u
p
to
1
4
3
.
6
5
MH
z
m
a
x
i
m
u
m
clo
c
k
f
r
eq
u
en
c
y
a
n
d
it
co
n
s
u
m
e
s
1
6
5
2
0
L
UT
s
th
at
r
ep
r
esen
ts
3
1
.
0
5
%
o
f
av
ailab
le
r
eso
u
r
ce
s
i
n
o
u
r
d
ev
ice
(
5
3
2
0
0
L
UT
s
)
an
d
1
3
4
5
r
eg
is
ter
s
b
its
co
r
r
esp
o
n
d
in
g
to
1
.
2
7
%
o
f
1
0
6
4
0
0
r
eg
is
ter
s
r
eso
u
r
ce
s
,
f
in
all
y
,
it
d
o
esn
’
t
e
x
p
lo
it
an
y
DSP
s
f
r
o
m
t
h
e
2
2
0
DSP
s
ex
itin
g
i
n
o
u
r
FP
GA
.
Mo
r
e
d
etails
o
f
th
e
s
tep
s
o
f
t
h
e
i
m
p
le
m
e
n
tat
io
n
i
n
p
latf
o
r
m
Vi
v
ad
o
I
n
teg
r
ated
D
esig
n
E
n
v
ir
o
n
m
e
n
t (
I
DE
)
ar
e
g
iv
e
n
i
n
th
e
w
o
r
k
[
2
0
]
.
T
h
e
s
i
m
u
latio
n
in
Mo
d
elSi
m
d
em
o
n
s
tr
ate
s
t
h
at
o
u
r
ar
ch
i
te
ctu
r
e
ca
n
co
m
p
u
te
3
2
×3
2
b
lo
ck
i
n
4
6
7
clo
ck
c
y
cle
s
an
d
ta
k
es
6
c
y
c
les
to
p
r
o
ce
s
s
o
n
e
P
r
o
ce
s
s
in
g
E
le
m
en
t
(
P
E
)
.
T
h
e
p
r
o
p
o
s
ed
ar
ch
itectu
r
e
ca
n
co
m
p
u
te
3
0
f
r
a
m
es
p
er
s
ec
o
n
d
(
f
p
s
)
f
o
r
4
K
r
eso
lu
tio
n
a
n
d
1
2
0
f
p
s
f
o
r
th
e
F
u
ll
HD
r
eso
lu
tio
n
.
I
n
f
ac
t,
f
o
r
th
e
4
K
(
3
8
4
0
×
2
1
6
0
p
ix
els)
r
eso
l
u
tio
n
s
,
to
ac
h
iev
e
3
0
f
r
am
e
s
p
er
s
ec
o
n
d
w
ith
4
:2
:0
s
a
m
p
lin
g
o
n
e
r
eq
u
ir
es
(
3
8
4
0
×2
1
6
0
×1
.
5
×3
0
)
i.e
3
7
3
.
2
4
MSa
m
p
le/s
a
s
t
h
r
o
u
g
h
p
u
t,
s
o
o
u
r
t
h
r
o
u
g
h
p
u
t
is
h
ig
h
er
t
h
a
n
t
h
i
s
v
alu
e.
Mo
r
eo
v
er
,
th
i
s
p
r
o
p
o
s
ed
h
ar
d
w
ar
e
ar
ch
itect
u
r
e
ca
n
b
e
i
m
p
r
o
v
ed
to
p
r
o
c
ess
m
o
r
e
f
r
am
es p
er
s
ec
o
n
d
b
y
ex
p
lo
itin
g
t
h
e
o
th
er
ex
i
s
ti
n
g
m
o
r
e
r
eso
u
r
ce
s
o
n
th
e
u
s
ed
d
ev
i
ce
.
T
h
e
p
r
o
ce
s
s
in
g
ti
m
e
o
f
h
ar
d
w
ar
e
ac
ce
ler
ate
w
a
s
co
m
p
u
te
d
b
y
co
u
n
ti
n
g
th
e
to
tal
n
u
m
b
er
o
f
clo
ck
c
y
cles
ta
k
e
n
to
d
eter
m
i
n
e
all
p
o
s
s
ib
le
o
u
tp
u
t
p
i
x
el
s
f
o
r
b
ask
etb
all
s
eq
u
e
n
ce
t
h
at
ta
k
es
in
to
co
n
s
id
er
atio
n
th
e
n
u
m
b
er
o
f
f
r
a
m
es,
r
eso
l
u
tio
n
a
n
d
all
p
o
s
s
ib
le
s
ize
b
lo
ck
.
T
h
e
w
h
o
le
ti
m
e
i
s
i
n
r
elatio
n
to
a
clo
ck
f
r
eq
u
en
c
y
t
h
at
is
eq
u
a
l to
1
4
3
.
6
5
MH
z.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
SS
N
:
2
0
8
8
-
8708
I
n
t J
E
lec
&
C
o
m
p
E
n
g
,
Vo
l.
9
,
No
.
6
,
Dec
em
b
er
2
0
1
9
:
5
0
8
4
-
509
4
5092
5
.
2
.
Co
m
pa
ri
s
o
n o
f
intr
a
predict
io
n
m
o
des
a
lg
o
rit
h
m
i
mp
le
m
ent
a
t
io
ns
o
n F
P
G
A
.
C
o
m
p
ar
is
o
n
o
u
r
w
o
r
k
o
n
FP
GA
i
m
p
le
m
e
n
tatio
n
o
f
H
E
VC
in
tr
a
p
r
ed
ictio
n
w
ith
t
h
e
p
r
o
p
o
s
ed
h
ar
d
w
ar
e
o
n
th
e
FP
G
A
i
n
[
2
1
]
,
it
h
as
less
ar
ea
an
d
h
i
g
h
f
r
eq
u
e
n
c
y
.
h
o
w
e
v
er
th
e
p
er
f
o
r
m
an
ce
t
h
e
its
ar
ch
itect
u
r
e
u
p
to
5
5
f
r
am
e
s
p
er
s
ec
o
n
d
o
n
l
y
f
o
r
r
eso
lu
tio
n
f
u
ll
HD
a
g
ain
s
t
o
u
r
u
p
to
r
ea
l
ti
m
e
p
r
o
ce
s
s
i
n
g
f
o
r
r
eso
lu
tio
n
4
K.
W
h
o
le
h
ar
d
w
a
r
e
ar
ch
itectu
r
e
f
o
r
d
ec
o
d
er
is
p
r
esen
ted
b
y
A
b
e
y
d
ee
r
a
e
t
al.
[
6
]
.
I
n
o
r
d
er
to
co
m
p
ar
e
o
u
r
p
er
f
o
r
m
an
ce
s
f
o
r
in
tr
a
p
r
ed
ictio
n
,
w
e
co
n
s
id
e
r
o
n
l
y
th
eir
in
tr
a
p
r
ed
ictio
n
r
esu
lt
s
p
r
esen
ted
in
T
ab
le
2
.
T
h
e
h
ar
d
w
ar
e
ac
ce
l
er
ato
r
is
i
m
p
le
m
en
ted
o
n
Xil
in
x
Z
y
n
q
7
0
4
5
w
ith
2
8
n
m
t
ec
h
n
o
lo
g
y
p
r
o
ce
s
s
.
T
h
e
co
n
s
u
m
ed
r
eso
u
r
ce
s
ar
ea
is
4
3
K
o
f
L
UT
s
a
n
d
2
2
K
R
eg
i
s
ter
s
.
T
h
e
9
4
B
R
A
Ms
(
B
L
OC
K
R
AM
)
ar
e
o
cc
u
p
ied
,
in
to
tall
y
1
8
Kb
its
f
r
o
m
t
h
e
en
tire
m
e
m
o
r
y
.
T
h
e
m
ax
i
m
u
m
f
r
eq
u
e
n
c
y
is
u
p
to
1
5
0
MH
z.
T
h
e
d
esig
n
ca
n
b
e
p
r
o
d
u
ce
d
in
th
e
av
er
ag
e
2
.
6
0
1
s
am
p
les
p
ix
e
ls
p
er
clo
ck
c
y
cle,
co
n
s
eq
u
en
tl
y
,
it
ca
n
b
e
u
p
r
ea
l
ti
m
e
p
r
o
ce
s
s
in
g
(
3
0
f
p
s
)
f
o
r
th
e
s
eq
u
en
ce
4
K.
T
ab
le
2
.
C
o
m
p
ar
is
o
n
o
f
S
y
n
t
h
esis
R
e
s
u
lts
f
o
r
FP
GA
I
m
p
le
m
en
tatio
n
O
u
r
w
o
rk
A
b
ay
d
eer
[6
]
Bi
ao
[7
]
A
mi
s
h
[8
]
A
n
g
el
h
ar
t
[9
]
J
i
an
g
[1
0
]
Kh
an
[1
1
]
H
as
a
n
[2
1
]
T
ech
n
o
l
o
g
y
(F
PG
A
)
2
8
n
m
A
r
t
i
x
-
7
2
8
n
m
A
rt
i
x
-
7
6
5
n
m
4
0
n
m
V
i
r
t
e
x
-
6
6
5
n
m
6
5
n
m
V
i
rt
ex
-
5
4
0
n
m
V
i
rt
ex
-
6
4
0
n
m
V
i
rt
ex
-
6
L
U
T
/
A
L
M
a
1
6
.
5
2
K
43K
14K
1
7
0
K
8
.
8
K
69k
1
4
0
k
4
.
4
k
Reg
i
s
t
ers
1
.
3
4
K
22K
5
.
5
K
1
1
0
K
1
3
.
K
1
.
1
K
Max
f
req
u
e
n
cy
1
4
3
Mh
z
1
5
0
MH
z
1
1
0
Mh
z
2
1
9
MH
z
1
1
0
MH
z
2
0
4
Mh
z
2
1
3
Mh
z
2
2
7
Mh
z
F
rame
ra
t
e
4
K
@
3
0
f
p
s
4
K
@
3
0
f
p
s
4
K
@
3
0
f
p
s
4
K
@
2
4
f
p
s
2
1
6
0
@
3
0
f
p
s
--
4
K
@
3
0
f
p
s
1
0
8
0
@
5
5
f
p
s
memo
ry
3
6
K
b
1
8
K
b
6K
--
9
6
K
b
1
4
8
k
b
1
5
0
k
RE
G
--
PU
s
ALL
ALL
ALL
ALL
ALL
ALL
ALL
ALL
a
L
o
o
k
-
u
p
T
a
b
l
e
(
L
U
T
)
f
o
r
X
i
l
i
n
x
F
P
G
A
a
n
d
(
A
d
a
p
t
i
v
e
L
o
g
i
c
M
o
d
u
l
e
(
A
L
M
)
f
o
r
d
e
v
i
c
e
s fr
o
m A
L
T
ER
A
I
N
C
.
T
h
e
r
elate
d
r
ef
er
en
ce
s
i
n
s
ta
t
e
o
f
ar
t
ar
e
s
h
o
w
n
i
n
T
ab
le
2
.
T
h
at
in
clu
d
es
l
is
t
s
k
e
y
p
e
r
f
o
r
m
an
ce
m
etr
ics
o
f
cir
cu
it
ar
ch
itect
u
r
e
s
f
o
r
in
tr
a
p
r
ed
ictio
n
as
FP
G
A
tec
h
n
o
lo
g
y
,
co
n
s
u
m
p
tio
n
o
f
L
UT
s
o
r
A
L
Ms
f
o
r
Xilin
x
d
ev
ices
o
r
f
o
r
AL
E
T
R
A
d
e
v
ices
r
esp
ec
ti
v
el
y
,
u
t
iliz
atio
n
o
f
r
e
g
is
ter
s
,
t
h
e
n
u
m
b
er
r
eso
lu
tio
n
f
r
a
m
es
ca
n
b
e
p
r
o
ce
s
s
in
g
p
er
s
ec
o
n
d
,
an
d
also
th
e
P
Us
s
ize
is
r
ea
ch
ed
.
O
u
r
w
o
r
k
an
d
all
r
e
f
er
e
n
ce
s
ar
e
t
h
e
s
a
m
e,
s
in
ce
all
v
ar
iab
le
P
Us
s
ize
f
o
r
in
tr
a
p
r
ed
ictio
n
ar
e
s
u
p
p
o
r
ted
.
Du
e
to
d
i
f
f
er
e
n
t
f
ab
r
icati
o
n
tech
n
o
lo
g
y
t
h
a
t
r
elate
d
th
e
m
a
x
i
m
u
m
f
r
eq
u
e
n
c
y
r
ea
ch
ed
f
o
r
ea
ch
h
ar
d
w
ar
e
d
esig
n
o
n
d
i
f
f
er
en
t
FP
GA
p
lat
f
o
r
m
s
s
h
o
u
ld
b
e
n
o
t
co
m
p
ar
ed
,
t
h
er
ef
o
r
e,
t
h
e
t
h
r
o
u
g
h
p
u
t
i
s
i
n
cl
u
d
ed
in
ca
lcu
l
atio
n
.
Fo
r
t
h
e
t
h
r
o
u
g
h
p
u
t,
al
l
h
ar
d
w
ar
e
ca
n
b
e
p
r
o
ce
s
s
in
r
ea
l
ti
m
e
4
K
v
id
eo
f
r
a
m
es
e
x
ce
p
t
i
n
[
8
]
,
w
h
ic
h
ca
n
b
e
p
r
o
ce
s
s
2
4
f
r
a
m
es
/s
.
B
o
th
t
h
i
s
w
o
r
k
an
d
[
6
]
h
av
e
t
h
e
s
a
m
e
m
e
m
o
r
y
o
cc
u
p
ied
in
t
h
e
d
esi
g
n
,
b
u
t
[
7
]
ac
h
iev
ed
m
in
i
m
u
m
s
ize
m
e
m
o
r
y
(
6
k
b
its
)
in
it
s
cir
cu
it,
h
o
w
e
v
er
,
in
[9
–
11]
h
a
v
e
a
h
i
g
h
e
r
r
eso
u
r
ce
o
f
m
e
m
o
r
y
,
co
n
s
eq
u
e
n
tl
y
,
th
e
y
d
esi
g
n
w
il
l
co
n
s
u
m
e
m
o
r
e
en
er
g
y
t
h
an
o
t
h
er
s
.
O
u
r
d
esi
g
n
r
ea
c
h
es
a
lo
w
er
f
r
eq
u
e
n
c
y
t
h
an
k
s
to
th
e
tech
n
o
lo
g
y
u
s
ed
an
d
also
d
u
e
to
th
e
p
r
o
p
o
s
e
ar
ch
itectu
r
al
d
esi
g
n
,
t
h
at
ca
n
lead
to
h
ig
h
er
b
it
r
ate.
So
,
o
u
r
p
r
o
p
o
s
ed
ar
ch
itectu
r
e
ca
n
ac
h
iev
e
h
ig
h
th
r
o
u
g
h
p
u
t
u
p
to
3
ti
m
e
s
co
m
p
ar
ed
w
it
h
t
h
e
b
est
r
es
u
lts
in
[
1
1
]
an
d
c
an
co
m
p
u
te
all
b
o
ck
s
s
izes
lik
e
in
[
1
1
]
,
w
ith
a
le
s
s
i
n
cr
ea
s
e
in
th
e
n
ee
d
ed
ar
ea
an
d
m
e
m
o
r
y
c
au
s
ed
b
y
t
h
e
r
ef
er
en
ce
s
a
m
p
le
d
ata.
6.
CO
NCLU
SI
O
N
In
th
i
s
p
ap
er
,
a
n
o
v
el
ef
f
icie
n
t
ac
ce
ler
ato
r
f
o
r
in
tr
a
p
r
e
d
ictio
n
o
f
HE
VC
is
p
r
esen
ted
,
o
u
r
h
ar
d
w
ar
e
d
esig
n
s
u
p
p
o
r
ted
all
m
o
d
es
f
o
r
in
tr
a
p
r
ed
ictio
n
in
HE
VC
,
a
n
d
it
i
n
clu
e
s
b
o
th
p
ar
allel
a
n
d
p
ip
elin
e
tech
n
iq
u
e
s
in
o
r
d
er
to
r
e
d
u
ce
th
e
co
m
p
u
tatio
n
al
co
m
p
le
x
it
y
an
d
to
h
a
v
e
h
i
g
h
ef
f
icie
n
t
t
h
r
o
u
g
h
p
u
t.
T
h
e
ar
ch
itectu
r
e
i
s
i
m
p
le
m
en
ted
i
n
th
e
X
ili
n
x
Z
y
n
q
-
7
0
0
0
A
r
ti
x
-
7
.
T
h
e
FP
GA
S
y
n
t
h
esi
s
r
esu
l
ts
s
h
o
w
th
a
t th
e
p
r
o
p
o
s
ed
h
ar
d
w
ar
e
ca
n
r
ea
ch
1
4
3
MH
z
o
f
th
e
m
ax
i
m
u
m
clo
ck
f
r
eq
u
en
c
y
a
n
d
ca
n
p
r
o
ce
s
s
in
r
ea
l
ti
m
e
th
e
h
i
g
h
r
eso
lu
tio
n
v
id
eo
4
K.
I
n
f
u
t
u
r
e
w
o
r
k
,
w
e
p
la
n
to
ad
d
th
e
b
lo
ck
s
m
o
r
e
co
m
p
le
x
s
u
c
h
as
i
n
ter
p
r
ed
ictio
n
a
n
d
en
tr
o
p
y
en
co
d
in
g
in
o
r
o
r
d
er
to
ac
h
iev
e
r
ea
l
ti
m
e
p
r
o
ce
s
s
in
g
f
o
r
co
-
d
esig
n
o
n
b
o
t
h
s
y
s
te
m
s
o
f
t
w
ar
e
a
n
d
h
ar
d
w
ar
e
in
s
tead
o
f
co
-
s
i
m
u
latio
n
w
it
h
d
ata
s
to
r
e
alr
ea
d
y
i
n
m
e
m
o
r
y
.
F
u
r
th
er
m
o
r
e,
d
etails
ab
o
u
t
p
o
w
er
esti
m
at
io
n
b
ased
o
n
s
y
s
te
m
o
u
ts
id
e
d
ev
i
ce
an
d
in
s
id
e
s
o
as
to
esti
m
at
e
th
e
co
n
s
u
m
p
tio
n
e
n
er
g
y
m
o
r
e
th
an
e
s
ti
m
ate
b
y
Xilin
x
to
o
l.
Evaluation Warning : The document was created with Spire.PDF for Python.
I
n
t J
E
lec
&
C
o
m
p
E
n
g
I
SS
N:
2
0
8
8
-
8708
A
n
en
erg
y
-
a
w
a
r
e
s
ystem
-
on
-
c
h
ip
a
r
ch
itectu
r
e
fo
r
in
tr
a
p
r
ed
ictio
n
in
HE
V
C
s
ta
n
d
a
r
d
(
A
b
d
ess
a
ma
d
el
A
n
s
a
r
i
)
5093
RE
F
E
R
E
NC
E
S
[1
]
J.
L
a
in
e
m
a
,
F
.
Bo
ss
e
n
,
W
.
-
J
.
Ha
n
,
J.
M
in
,
a
n
d
K.
Ug
u
r,
“
In
tra
c
o
d
i
n
g
o
f
th
e
HE
V
C
sta
n
d
a
rd
,
”
IEE
E
T
ra
n
s.
Circ
u
it
s S
y
st.
Vi
d
e
o
T
e
c
h
n
o
l
.
,
v
o
l
.
2
2
,
n
o
.
1
2
,
p
p
.
1
7
9
2
–
1
8
0
1
,
2
0
1
2
.
[2
]
F
.
Bo
ss
e
n
,
B
.
Bro
ss
,
K.
S
u
h
rin
g
,
a
n
d
D.
F
ly
n
n
,
“
HEVC
c
o
m
p
lex
it
y
a
n
d
i
m
p
le
m
e
n
tatio
n
a
n
a
ly
sis,”
IEE
E
T
ra
n
s.
Circ
u
it
s S
y
st.
Vi
d
e
o
T
e
c
h
n
o
l
.
,
v
o
l
.
2
2
,
n
o
.
1
2
,
p
p
.
1
6
8
5
–
1
6
9
6
,
2
0
1
2
.
[3
]
B.
P
e
n
g
,
T
.
W
a
n
g
,
X
.
Jin
,
a
n
d
C.
W
a
n
g
,
“
A
n
A
c
c
e
lera
ti
n
g
S
o
lu
ti
o
n
f
o
r
-
Bo
d
y
M
ON
D S
i
m
u
latio
n
w
it
h
F
P
G
A
-
S
o
C,
”
In
t.
J
.
Rec
o
n
fi
g
u
r
a
b
le C
o
mp
u
t.
,
v
o
l.
2
0
1
6
,
2
0
1
6
.
[4
]
“
Z
y
n
q
-
7
0
0
0
A
ll
P
r
o
g
ra
m
m
a
b
le
S
o
C,
”
3
0
-
Ja
n
-
2
0
1
7
.
[
On
li
n
e
].
Av
a
il
a
b
le:
h
tt
p
:/
/w
ww
.
x
il
in
x
.
c
o
m
/
p
ro
d
u
c
ts/sil
ico
n
-
d
e
v
ice
s/so
c
/z
y
n
q
-
7
0
0
0
.
h
tm
l.
[
A
c
c
e
ss
e
d
:
3
0
-
Ja
n
-
2
0
1
7
]
.
[5
]
L
.
H.
Cro
c
k
e
tt
,
R.
A
.
El
li
o
t,
M
.
A
.
En
d
e
rw
it
z
,
a
n
d
R.
W
.
S
tew
a
rt
,
T
h
e
Z
y
n
q
Bo
o
k
:
Emb
e
d
d
e
d
Pro
c
e
ss
in
g
wi
th
th
e
Arm Co
rte
x
-
A9
o
n
t
h
e
Xi
li
n
x
Z
y
n
q
-
7
0
0
0
Al
l
Pro
g
r
a
mm
a
b
le S
o
c
.
S
t
ra
th
c
ly
d
e
A
c
a
d
e
m
ic M
e
d
ia,
2
0
1
4
.
[6
]
M
.
A
b
e
y
d
e
e
ra
,
M
.
Ka
ru
n
a
ra
tn
e
,
G
.
Ka
ru
n
a
ra
tn
e
,
K.
De
S
il
v
a
,
a
n
d
A
.
P
a
sq
u
a
l,
“
4
K
re
a
l
-
ti
m
e
HEVC
d
e
c
o
d
e
r
o
n
a
n
F
P
GA
,
”
IEE
E
T
ra
n
s.
Circ
u
it
s
S
y
st.
Vi
d
e
o
T
e
c
h
n
o
l
.
,
v
o
l.
2
6
,
n
o
.
1
,
p
p
.
2
3
6
–
2
4
9
,
2
0
1
6
.
[7
]
B.
M
in
,
Z
.
Xu
,
a
n
d
R.
C.
Ch
e
u
n
g
,
“
A
F
u
ll
y
P
ip
e
li
n
e
d
Ha
rd
w
a
re
A
rc
h
it
e
c
tu
re
f
o
r
In
tra
P
re
d
ictio
n
o
f
HEV
C,
”
IEE
E
T
ra
n
s.
C
irc
u
it
s S
y
st.
Vi
d
e
o
T
e
c
h
n
o
l.
,
2
0
1
6
.
[8
]
F
.
Am
ish
a
n
d
E
.
-
B.
B
o
u
re
n
n
a
n
e
,
“
F
u
ll
y
p
ip
e
li
n
e
d
re
a
l
ti
m
e
h
a
rd
w
a
r
e
so
lu
ti
o
n
f
o
r
Hig
h
Ef
f
ici
e
n
c
y
V
id
e
o
Co
d
in
g
(HEV
C)
i
n
tra p
re
d
ictio
n
,
”
J
.
S
y
st.
Arc
h
it
.
,
v
o
l.
6
4
,
p
p
.
1
3
3
–
1
4
7
,
2
0
1
6
.
[9
]
D.
En
g
e
lh
a
rd
t,
J.
M
o
ll
e
r,
J.
Ha
h
l
b
e
c
k
,
a
n
d
B.
S
tab
e
r
n
a
c
k
,
“
F
P
G
A
im
p
le
m
e
n
tatio
n
o
f
a
f
u
ll
HD
re
a
l
-
ti
m
e
HEV
C
m
a
in
p
ro
f
il
e
d
e
c
o
d
e
r,
”
IEE
E
T
ra
n
s.
Co
n
s
u
m.
El
e
c
tro
n
.
,
v
o
l.
6
0
,
n
o
.
3
,
p
p
.
4
7
6
–
4
8
4
,
2
0
1
4
.
[1
0
]
W
.
Jia
n
g
,
H.
M
a
,
a
n
d
Y.
Ch
e
n
,
“
G
r
a
d
ien
t
b
a
se
d
f
a
st
m
o
d
e
d
e
c
isio
n
a
lg
o
ri
th
m
f
o
r
in
tra
p
re
d
icti
o
n
in
H
EV
C
,
”
in
Co
n
su
me
r
El
e
c
tro
n
ics
,
Co
mm
u
n
ica
ti
o
n
s
a
n
d
Ne
two
rk
s
(
CECNet
),
2
0
1
2
2
n
d
In
ter
n
a
ti
o
n
a
l
Co
n
f
e
re
n
c
e
o
n
,
p
p
.
1
8
3
6
–
1
8
4
0
,
2
0
1
2
.
[1
1
]
M
.
U.
K.
Kh
a
n
,
M
.
S
h
a
f
iq
u
e
,
M
.
G
re
ll
e
rt,
a
n
d
J.
He
n
k
e
l,
“
Ha
rd
wa
re
-
so
ft
wa
re
c
o
ll
a
b
o
ra
ti
v
e
c
o
mp
l
e
x
it
y
re
d
u
c
ti
o
n
sc
h
e
me
fo
r
th
e
e
me
rg
in
g
HEVC
i
n
tra
e
n
c
o
d
e
r
,
”
in
De
sig
n
,
A
u
to
m
a
ti
o
n
&
T
e
st
in
Eu
r
o
p
e
Co
n
f
e
re
n
c
e
&
Ex
h
ib
it
io
n
(DA
T
E)
, 2
0
1
3
,
p
p
.
1
2
5
–
1
2
8
,
2
0
1
3
.
[1
2
]
C.
-
T
.
Hu
a
n
g
,
M
.
T
ik
e
k
a
r,
a
n
d
A
.
P
.
Ch
a
n
d
ra
k
a
sa
n
,
“
M
e
m
o
r
y
-
h
iera
r
c
h
ica
l
a
n
d
m
o
d
e
-
a
d
a
p
ti
v
e
HEV
C
in
tra
p
re
d
ictio
n
a
rc
h
it
e
c
tu
re
f
o
r
q
u
a
d
fu
ll
HD
v
id
e
o
d
e
c
o
d
in
g
,
”
IEE
E
T
r
a
n
s.
Ver
y
L
a
rg
e
S
c
a
le In
teg
r.
V
L
S
I
S
y
st.
,
v
o
l.
2
2
,
n
o
.
7
,
p
p
.
1
5
1
5
–
1
5
2
5
,
2
0
1
4
.
[1
3
]
E.
Ka
lali,
Y.
A
d
ib
e
ll
i,
a
n
d
I.
H
a
m
z
a
o
g
lu
,
“
A
lo
w
e
n
e
rg
y
in
tra
p
re
d
icti
o
n
h
a
r
d
w
a
re
f
o
r
h
ig
h
e
ff
icie
n
c
y
v
id
e
o
c
o
d
in
g
,
”
J
.
Re
a
l
-
T
ime
Ima
g
e
Pro
c
e
ss
.
,
p
p
.
1
–
1
4
,
2
0
1
4
.
[1
4
]
M
.
Ka
m
m
o
u
n
,
A
.
B.
A
ti
tallah
,
a
n
d
N.
M
a
sm
o
u
d
i,
“
An
o
p
ti
mize
d
h
a
rd
wa
re
a
rc
h
i
tec
tu
re
fo
r
in
tra
p
re
d
ictio
n
f
o
r
HEVC
,
”
in
Im
a
g
e
P
ro
c
e
ss
in
g
,
Ap
p
li
c
a
ti
o
n
s
a
n
d
S
y
ste
m
s
Co
n
f
e
r
e
n
c
e
(IP
A
S
),
2
0
1
4
Fi
rs
t
In
ter
n
a
ti
o
n
a
l
,
p
p
.
1
–
5
,
2
0
1
4
.
[1
5
]
M
.
Je
o
n
a
n
d
B.
-
D.
L
e
e
,
“
T
o
wa
rd
Co
n
ten
t
-
A
w
a
r
e
V
id
e
o
P
a
rti
t
io
n
i
n
g
M
e
th
o
d
s
f
o
r
Distrib
u
ted
HEV
C
V
id
e
o
En
c
o
d
in
g
,
”
In
t
.
J
.
El
e
c
tr.
C
o
mp
u
t
.
En
g
.
IJ
ECE
,
v
o
l.
5
,
n
o
.
3
,
p
p
.
5
6
9
–
5
7
8
,
2
0
1
5
.
[1
6
]
G
.
J.
S
u
ll
iv
a
n
,
J.
Oh
m
,
W
.
-
J.
Ha
n
,
a
n
d
T
.
W
ieg
a
n
d
,
“
Ov
e
r
v
ie
w
o
f
th
e
h
ig
h
e
ff
icie
n
c
y
v
id
e
o
c
o
d
in
g
(HEV
C
)
sta
n
d
a
rd
,
”
IEE
E
T
r
a
n
s.
Circ
u
it
s S
y
st.
Vi
d
e
o
T
e
c
h
n
o
l.
,
v
o
l.
2
2
,
n
o
.
1
2
,
p
p
.
1
6
4
9
–
1
6
6
8
,
2
0
1
2
.
[1
7
]
I.
-
K.
Kim
,
K.
M
c
Ca
n
n
,
K.
S
u
g
i
m
o
to
,
B.
Br
o
ss
,
a
n
d
W
.
-
J.
Ha
n
,
“
H
m
9
:
Hig
h
e
f
f
ici
e
n
c
y
v
id
e
o
c
o
d
i
n
g
(HEV
C)
tes
t
m
o
d
e
l
9
e
n
c
o
d
e
r
d
e
sc
rip
ti
o
n
,
”
i
n
Pro
c
.
9
t
h
J
CT
-
VC
M
e
e
ti
n
g
,
p
p
.
6
–
11
,
2
0
1
2
.
[1
8
]
R.
P
.
P
.
S
i
n
g
h
,
P
.
Ku
m
a
r,
a
n
d
B.
S
in
g
h
,
“
P
e
rf
o
rm
a
n
c
e
a
n
a
l
y
si
s
o
f
3
2
-
b
it
a
rra
y
m
u
lt
ip
li
e
r
w
it
h
a
c
a
rry
sa
v
e
a
d
d
e
r
a
n
d
w
it
h
a
c
a
rry
-
lo
o
k
-
a
h
e
a
d
a
d
d
e
r,
”
In
t.
J
.
Rec
e
n
t
T
re
n
d
s E
n
g
.
,
v
o
l
.
2
,
n
o
.
6
,
p
p
.
8
3
–
8
6
,
2
0
0
9
.
[1
9
]
N.
M
.
Zab
id
i
a
n
d
A
.
A
.
-
H.
A
b
R
a
h
m
a
n
,
“
V
L
S
I
De
si
g
n
o
f
a
F
a
st
P
ip
e
li
n
e
d
8
x
8
Disc
re
te
Co
sin
e
T
ra
n
sf
o
rm
,
”
In
t.
J
.
El
e
c
tr.
Co
mp
u
t.
E
n
g
.
IJ
ECE
,
v
o
l.
7
,
n
o
.
3
,
p
p
.
1
4
3
0
–
1
4
3
5
,
2
0
1
7
.
[2
0
]
A
.
Ra
n
i
a
n
d
N.
G
ro
v
e
r,
“
A
n
En
h
a
n
c
e
d
F
P
G
A
Ba
s
e
d
A
s
y
n
c
h
ro
n
o
u
s
M
icr
o
p
r
o
c
e
ss
o
r
De
sig
n
Us
in
g
V
IV
A
DO
a
n
d
IS
IM
,
”
B
u
ll
.
El
e
c
tr.
En
g
.
I
n
fo
rm.
,
v
o
l.
7
,
n
o
.
2
,
p
p
.
1
9
9
–
2
0
8
,
2
0
1
8
.
[2
1
]
H.
A
z
g
in
,
A
.
C.
M
e
rt,
E
.
Ka
lali,
a
n
d
I
.
Ha
m
z
a
o
g
lu
,
“
An
e
ff
icie
n
t
F
PGA
imp
lem
e
n
t
a
ti
o
n
o
f
HEVC
in
tra
p
re
d
icti
o
n
,
”
in
Co
n
su
me
r E
lec
tro
n
ics
(
ICCE)
,
2
0
1
8
IE
EE
I
n
tern
a
ti
o
n
a
l
C
o
n
f
e
re
n
c
e
o
n
,
p
p
.
1
–
5
,
2
0
1
8
.
B
I
O
G
RAP
H
I
E
S
O
F
AUTH
O
RS
E
l
An
s
a
r
i
A
b
d
e
ss
a
m
a
d
wa
s
b
o
rn
i
n
M
o
r
o
c
c
o
in
.
In
1
9
8
7
.
h
e
re
c
e
iv
e
d
h
is
m
a
st
e
r
f
ro
m
th
e
f
a
c
u
lt
y
o
f
sc
ien
c
e
a
n
d
tec
h
n
o
l
o
g
y
(F
S
T
)
o
f
F
e
z
M
o
ro
c
c
o
,
in
2
0
1
2
,
h
e
j
o
i
n
e
d
t
h
e
tea
m
m
icro
e
le
c
tro
n
ics
a
n
d
e
m
b
e
d
e
e
d
s
y
ste
m
s
o
f
th
e
lab
o
ra
to
r
y
o
f
e
n
e
rg
ies
a
n
d
s
m
a
rt
s
y
ste
m
.
His
re
se
a
rc
h
in
tere
sts
in
c
lu
d
e
v
id
e
o
c
o
d
in
g
a
n
d
im
p
le
m
e
n
tatio
n
o
f
v
id
e
o
sta
n
d
a
rd
o
n
e
m
b
e
d
d
e
d
p
ro
c
e
ss
o
r
a
n
d
V
L
S
I
a
rc
h
it
e
c
tu
re
.
Evaluation Warning : The document was created with Spire.PDF for Python.