Information Technology Reference
In-Depth Information
Algorithm 1.
Computation of the number of collisions in a chain of vehicles.
for all
comb
in
Combinations
do
for all
d
in
RangeOfDistances
do
for
i
=1
to
N
do
p
i
=
f
(
p
i−
1
, comb, d, i, veloc, reactTime
)
end for
for
j
=0
to
N
do
Π
j
=
P
N
(1
,
(
N
+1)(
N
+2)
2
− j
)
end for
N
N
acc
=
j · Π
j
j
=0
end for
end for
Ta b l e 2 .
Resulting programs with different parallelized tasks. X means that the corresponding
parallelization takes place.
Program
A
B
C
Program 1
Program 2
×
Program 3
×
Program 4
×
× ×
Program 5
Program 6
×
×
Program 7
× ×
Program 8
× × ×
For the sake of clarity, we will parallelize the following tasks:
-
A: Vector-Matrix multiplication.
-
B: Average inter-vehicle distance variation.
-
C: Technology penetration rate variation.
Next, we will combine the different parallelized tasks (see Table 2) and execute the
resulting programs in order to assess the actual improvement obtained from each one.
6R su s
In this section we summarize the results obtained by executing the programs shown in
Table 2 in a node of the Arabi cluster. We have used
2
,
4
and
8
processors in order to
assess the improvement on the execution time achieved by each one.
The parameters used to execute the model are the following:
-
CCA penetration rate:
0%
−
50%
,in
10%
steps.
-
Average inter-vehicle distance:
6
−
70
m
,in
1
meter steps.
-
Number of vehicles:
20
vehicles.
-
Average velocity:
33
m/s
.
-
Average driver's reaction time:
1
s
.
Search WWH ::
Custom Search