DocumentCode :
2142457
Title :
Low power consumption scheduling based on software Fault-tolerance
Author :
Tianwen Yao ; Haifang Zhou ; Minquan Fang ; Haibin Hu
Author_Institution :
Coll. of Comput. Sci., Nat. Univ. of Defense Technol., Changsha, China
fYear :
2013
fDate :
23-25 July 2013
Firstpage :
1788
Lastpage :
1793
Abstract :
The space computer puts forward high demands on the performance. Therefore, the high-performance digital signal processors are increasingly used in the space computer. However, the single particle effects caused by the cosmic radiation make the reliability of the space computer become a huge challenge. The COTS DSP chip has a huge advantage compared to the antiradiation DSP chip in performance, price, size and weight. The software implemented fault-tolerance technique can protect the program, but degrade the system performance and increase the power consumption. According to the DSP structural characteristics and in the premise of not reducing the detecting error ratio, this paper proposes an instruction scheduling method for the low power consumption, to reduce the overheads in terms of the performance and the energy incurred by the fault-tolerance technique.
Keywords :
aerospace computing; digital signal processing chips; power aware computing; scheduling; software fault tolerance; COTS DSP chip; DSP structural characteristics; anti-radiation DSP chip; commercial-off-the-shelf DSP chip; cosmic radiation; high-performance digital signal processors; instruction scheduling method; low power consumption scheduling; overhead reduction; single particle effects; software fault-tolerance; space computer; Digital signal processing; Energy consumption; Fault tolerance; Fault tolerant systems; Optimization; Registers; Software; DSP; instruction scheduling; power optimization; software fault-tolerant;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Natural Computation (ICNC), 2013 Ninth International Conference on
Conference_Location :
Shenyang
Type :
conf
DOI :
10.1109/ICNC.2013.6818273
Filename :
6818273
Link To Document :
بازگشت