Fault Tolerant Issues in the BTeV Trigger, Introduction: The BTeV Trigger.

Butler, J. N. ; Appel, J. A. ; Brennan, T. ; Gottschalk, E. ; McBride, P. ; Petravick, D. ; Pordes, R. ; Votava, M. ; Iyer, K. ; Kalbarczyk, Z. T.

Technical Information Center Oak Ridge Tennessee

关键词: Accelerators;    Fault tolerance analysis;    Algorithms;    Data base management;    Microprocessors;   

RP-ID  :  DE141131224

学科分类：工程和技术（综合）

美国|英语

来源: National Technical Reports Library

【 摘 要 】

The BTeV trigger performs sophisticated computations using large ensembles of FPGAs, DSPs, and conventional microprocessors. This system will have between 5,000 and 10,000 computing elements and many networks and data switches. While much attention has been devoted to developing efficient algorithms, the need for fault-tolerant, fault-adaptive, and flexible techniques and software to manage this huge computing platform has been identified as one of the most challenging aspects of this project. We describe the problem and offer an approach to solving it based on a distributed, hierarchical fault management system.