Hardware Reference
In-Depth Information
Table 5.11
Energy consumption with no load (10 s)
Energy
consumption
(Ws)
Comparison with
the performance
governor (%)
Governor
Core condition
Idle reduction
5.2
26
75 MHz with sleep instruction × 1
CPU hot remove × 3
On demand
6.3
31.5
75 MHz with sleep instruction × 4
Conservative
6.3
31.5
75 MHz with sleep instruction × 4
Powersave
6.2
31
75 MHz with sleep instruction × 4
Performance
20
100
600 MHz with sleep instruction × 4
Idle Reduction
ondemand
conservative
powersave
performance
20
15
10
5
0
0
2
4
6
8
10
12
Elapsed Time [s]
Fig. 5.3
Comparison of energy consumption
Table 5.12
Power consumption with mixed loads
Energy
consumption
(Ws)
Comparison with
performance
governor (%)
Execution
time (s)
Average power
consumption (W)
Governor
Idle reduction
103.3
89
52
2
On demand
121.6
104.7
52
2.3
Conservative
111.1
95.7
73
1.5
Powersave
126.2
108.7
187
0.7
Performance
116.1
100
44
2.6
powersave had the second lowest. Idle reduction was able to reduce energy con-
sumption by 16% compared to the powersave governor.
Table
5.12
lists the energy consumption of the two-thread RAYTRACE bench-
mark; two cores have high loads (executing threads) and the other two cores have no
load (sleep or CPU hot remove). Idle reduction has the lowest energy consumption
(103.3 Ws), and the conservative governor has the second lowest (111.1 Ws). Idle
reduction was able to reduce power consumption by 7% compared to the conserva-
tive governor.
Search WWH ::
Custom Search