2009.3: Algorithmic Based Fault Tolerance Applied to High Performance Computing
2009.3: George Bosilca, Remi Delmas, Jack Dongarra and Julien Langou (2009) Algorithmic Based Fault Tolerance Applied to High Performance Computing.
Full text available as:
|PDF - Requires a PDF viewer such as GSview, Xpdf or Adobe Acrobat Reader|
We present a new approach to fault tolerance for High Performance Computing system. Our approach is based on a careful adaptation of the Algorithmic Based Fault Tolerance technique (Huang and Abraham, 1984) to the need of parallel distributed computation. We obtain a strongly scalable mechanism for fault tolerance. We can also detect and correct errors (bit-flip) on the fly of a computation. To assess the viability of our approach, we have developed a fault tolerant matrixmatrix multiplication subroutine and we propose some models to predict its running time. Our parallel fault-tolerant matrix-matrix multiplication scores 1.4 TFLOPS on 484 processors (cluster jacquard.nersc.gov) and returns a correct result while one process failure has happened. This represents 65% of the machine peak efficiency and less than 12% overhead with respect to the fastest failure-free implementation. We predict (and have observed) that, as we increase the processor count, the overhead of the fault tolerance drops significantly.
|Item Type:||MIMS Preprint|
Appears also as Technical Report UT-CS-08-620, Department of Computer Science, University of Tennessee, Knoxville, TN, USA, June 2008 and as LAPACK Working Note 205"
|Subjects:||MSC 2000 > 65 Numerical analysis|
MSC 2000 > 68 Computer science
|Deposited By:||Ms Lucy van Russelt|
|Deposited On:||13 January 2009|