VLSI Digital Signal Processing Systems

Iteration Bound
Lan-Da Van (), Ph. D.
Department of Computer Science
National Chiao Tung University
Taiwan, R.O.C.
Fall, 2010
ldvan@cs.nctu.edu.tw
http://www.cs.nctu.tw/~ldvan/
VLSI Digital Signal Processing Systems
Lan-Da Van VLSI-DSP-2-2
Outline
Introduction
Data Flow Graph (DFG) Representations
Loop Bound and Iteration Bound
Compute the Iteration Bound
Longest Path Matrix Algorithm (LPM)
Minimum Cycle Mean Method (MCM)
Conclusion
VLSI Digital Signal Processing Systems
Lan-Da Van VLSI-DSP-2-3
Introduction
Many DSP algorithms such as recursive and
adaptive digital filters contain feedback loops,
which impose an inherent fundamental lower
bound on the achievable iteration or sample period.
Iteration bound = maximum loop bound
Clock period = critical path period = cycle time =
1/clock rate
Sample rate = throughput rate
Impossible to achieve an iteration bound less than
the theoretical iteration bound with infinite
processors
VLSI Digital Signal Processing Systems
Lan-Da Van VLSI-DSP-2-4
Outline
Introduction
Data Flow Graph (DFG) Representations
Loop Bound and Iteration Bound
Compute the Iteration Bound
Longest Path Matrix Algorithm (LPM)
Minimum Cycle Mean Method (MCM)
Conclusion
VLSI Digital Signal Processing Systems
Lan-Da Van VLSI-DSP-2-5
DSP Program: Example
) ( ) 1 ( ) ( n x n ay n y + =
= 0 for to n
Data Flow Graph
k k
B A : Period Iteration - Intra
1
: Period Iteration - Inter
+

k k
A B
VLSI Digital Signal Processing Systems
Lan-Da Van VLSI-DSP-2-6
Loop Bounds in DFG
Critical Path: The path with the
longest computation time among
all paths that contain zero delays
Loop: Directed path that begins
and ends at the same node
Loop Bound=t
l
/w
l
, where t
l
is the
loop computation time and w
l
is the
number of delays in the loop
Critical Loop: the loops in which
has maximum loop bound.
Iteration Bound: maximum loop
bound, i.e., a fundamental limit for
recursive algorithms
Loop bounds : 4/2 u.t. (max), 5/3 u.t. , 5/4 u.t.
VLSI Digital Signal Processing Systems
Lan-Da Van VLSI-DSP-2-7
Iteration bound
Definition:
(b) Iteration bound= Max { 6/2 , 11/1 }
= 11 u.t.
T
} { max
l
l
L l
w
t
T
e

=
(a) Iteration bound= 6/2 = 3 u.t.
VLSI Digital Signal Processing Systems
Lan-Da Van VLSI-DSP-2-8
Outline
Introduction
Data Flow Graph (DFG) Representations
Loop Bound and Iteration Bound
Compute the Iteration Bound
Longest Path Matrix Algorithm (LPM)
Minimum Cycle Mean Method (MCM)
Conclusion
VLSI Digital Signal Processing Systems
Lan-Da Van VLSI-DSP-2-9
Longest Path Matrix (LPM) Algorithm
A series of matrix is constructed, and the iteration
bound is found by examining the diagonal elements
of the matrices.
d : # of delays
Compute L
(1)
~L
(m)
, m=1, 2, 3, , d
:The longest computation time of all paths from delay
element d
i
to delay element d
j
that pass through exactly m-1
delays, where the delay d
i
and delay d
j
are not included for m-1
delays.
If no path exists, then the value of l
i,j
(m)
equals -1.
) (
,
m
j i
l
VLSI Digital Signal Processing Systems
Lan-Da Van VLSI-DSP-2-10
A DFG with Three Loops Using LPM
(1/3)
(
(
(
(

=
1 1 1 5
0 1 1 5
1 0 1 4
1 1 0 1
) 1 (
L
VLSI Digital Signal Processing Systems
Lan-Da Van VLSI-DSP-2-11
A DFG with Three Loops Using LPM
(2/3)
) , 1 ( max
) (
,
) 1 (
,
) 1 (
,
m
j k k i
K k
m
j i
l l l + =
e
+
(
(
(
(

=
1 1 5 1
1 1 5 5
0 1 4 5
1 0 1 4
) 2 (
L
(
(
(
(

(
(
(
(

=
1 1 1 5
0 1 1 5
1 0 1 4
1 1 0 1
1 1 1 5
0 1 1 5
1 0 1 4
1 1 0 1
) 2 (
L
VLSI Digital Signal Processing Systems
Lan-Da Van VLSI-DSP-2-12
A DFG with Three Loops Using LPM
(3/3)
} { max
) (
,
} ,..., 2 , 1 { ,
m
l
T
m
i i
d m i e

=
2 }
4
5
,
4
5
,
4
8
,
4
8
,
3
5
,
3
5
,
3
5
,
2
4
,
2
4
{ max
} ,..., 2 , 1 { ,
= =
e

d m i
T
(
(
(
(

=
1 5 1 9
1 5 5 9
1 4 5 8
0 1 4 5
) 3 (
L
(
(
(
(

=
5 1 9 10
5 5 9 10
4 5 8 9
1 4 5 8
) 4 (
L
VLSI Digital Signal Processing Systems
Lan-Da Van VLSI-DSP-2-13
A Filter Using LPM
(

=
8 8
4 4
) 1 (
L
(

=
16 16
12 12
) 2 (
L
8 }
2
16
,
2
12
,
1
8
,
1
4
max{ = =

T
VLSI Digital Signal Processing Systems
Lan-Da Van VLSI-DSP-2-14
1. Construct the new graph G
d and
Gd
Transform from original DFG G
Decide the # of nodes from the # of delays in G.
Decide the weight of each edge
2. Compute the minimum cycle mean
Construct the series of d+1 vectors f
(m)
, m=0, 1, 2, ,d
An arbitrary reference node is chosen in Gd (called this
node s). The initial vector f
(0)
is formed by setting f
(0)
(s)=0
and setting the remaining nodes of f
(0)
to infinity.
find the min cycle mean
Minimum Cycle Mean (MCM) Method
VLSI Digital Signal Processing Systems
Lan-Da Van VLSI-DSP-2-15
Cycle mean= Average length of the edge in c
(Cycle = Loop )
Longest path length
path that passes through no delays
longest : two loops that contain D
a
and D
b
max { 6,4 } = 6
cycle mean = 6/2=3
A DFG with Three Loops Using MCM
(1/4)
VLSI Digital Signal Processing Systems
Lan-Da Van VLSI-DSP-2-16
delay => node
longest path length (computation time) =>weight
w(i,j)
If no zero-delay path exists from delay d
i
to delay d
j
,
then the edge i -> j does not exist in Gd.
A DFG with Three Loops Using MCM
(2/4)
VLSI Digital Signal Processing Systems
Lan-Da Van VLSI-DSP-2-17
we will find d+1 vectors , f
(m)
m=0,1,, d
A DFG with Three Loops Using MCM
(3/4)
(
(
(
(

=
0
) 0 (
f
(
(
(
(

=
0
) 1 (
f
(
(
(
(

=
0
4
) 2 (
f
(
(
(
(

=
0
4
5
) 3 (
f
(
(
(
(

=
4
5
8
) 4 (
f
)) , ( ) ( ( min ) (
) 1 ( ) (
j i w i f j f
m
I i
m
+ =

e
VLSI Digital Signal Processing Systems
Lan-Da Van VLSI-DSP-2-18
A DFG with Three Loops Using MCM
(4/4)
2 } , 2 , 1 , 2 min{ = =

T
))
) ( ) (
( max ( min
) ( ) (
} 1 ,... 2 , 1 , 0 { } ,..., 2 , 1 {
m d
i f i f
T
m d
d m d i

=
e e

VLSI Digital Signal Processing Systems

Lan-Da Van VLSI-DSP-2-19
A Filter Using MCM
8 } 6 , 8 min{ = =

T
(

=
12
12
) 2 (
f
(

=
4
4
) 1 (
f
(

=
0
) 0 (
f
VLSI Digital Signal Processing Systems
Lan-Da Van VLSI-DSP-2-20
Conclusion
When the DFG is recursive, the iteration bound is
the fundamental limit on the minimum sample
period of a hardware implementation of the DSP
program.
Two algorithms to compute iteration bound, LPM
and MCM, were explored.