5866 search results - page 181 / 1174 » Designing Accelerator-Based Distributed Systems for High Per... |
161
click to vote
IPPS
2007
IEEE
16 years 24 days ago
2007
IEEE
There are two popular parallel programming paradigms available to high performance computing users such as engineering and physics professionals: message passing and distributed s...
173
click to vote
ICPPW
2006
IEEE
16 years 16 days ago
2006
IEEE
Abstract— Most image processing applications are characterized by computation-intensive operations, and high memory and performance requirements. Parallelized implementation on s...
151
click to vote
MSS
2005
IEEE
16 years 2 days ago
2005
IEEE
We present a policy-based architecture STEPS for lifecycle management (LCM) in a mass scale distributed file system. The STEPS architecture is designed in the context of IBM’s ...
162
click to vote
IPPS
1999
IEEE
15 years 10 months ago
1999
IEEE
Our study of a large set of scientific applications over the past three years indicates that the processing for multidimensional datasets is often highly stylized. The basic proce...
191
click to vote
CLUSTER
2006
IEEE
16 years 17 days ago
2006
IEEE
Most of today‘s HPC systems employ a single head node for control, which represents a single point of failure as it interrupts an entire HPC system upon failure. Furthermore, it...