DocumentCode :
2155806
Title :
Testing and fault tolerance properties in a class of sorting networks
Author :
Salloum, Salam ; Gala, Murali
Author_Institution :
Houston-Victoria Univ., Victoria, TX, USA
Volume :
2
fYear :
1997
fDate :
20-22 Aug 1997
Firstpage :
938
Abstract :
Sorting networks are an important application for parallel processing and also can be used as interconnection networks with fast and simple self routing algorithm. The odd-even transposition sorting networks are an important class of sorting networks because of the ease of hardware implementation. The paper investigates testing and fault tolerant properties of such networks for single stuck at crossing (stuck at X) and stuck at straight (stuck at T) faults in the comparators. We have proved that a single test vector is sufficient to detect all single and multiple stuck at T faults. For stuck at X faults, we have shown that the networks are 1-fault tolerant for large number of comparators. Analysis of several examples support the conjecture that these networks are 1-fault tolerant for all internal comparators
Keywords :
computer networks; fault location; fault tolerant computing; parallel algorithms; reliability; sorting; 1-fault tolerant; comparators; fault tolerance properties; hardware implementation; interconnection networks; multiple stuck at T faults; odd-even transposition sorting networks; parallel processing; self routing algorithm; single stuck at crossing faults; single test vector; stuck at X faults; stuck at straight faults; testing; Digital signal processing; Fault tolerance; Hardware; Instruments; Intelligent networks; Multiprocessor interconnection networks; Parallel processing; Rivers; Sorting; Testing;
fLanguage :
English
Publisher :
ieee
Conference_Titel :
Communications, Computers and Signal Processing, 1997. 10 Years PACRIM 1987-1997 - Networking the Pacific Rim. 1997 IEEE Pacific Rim Conference on
Conference_Location :
Victoria, BC
Print_ISBN :
0-7803-3905-3
Type :
conf
DOI :
10.1109/PACRIM.1997.620414
Filename :
620414
Link To Document :
بازگشت