Connectionist learning procedures
Artificial Intelligence
Practical prefetching techniques for multiprocessor file systems
Distributed and Parallel Databases - Selected papers from the first international conference on parallel and distributed information systems
Unstructured spectral element methods for simulation of turbulent flows
Journal of Computational Physics
Informed prefetching and caching
SOSP '95 Proceedings of the fifteenth ACM symposium on Operating systems principles
PPFS: a high performance portable parallel file system
ICS '95 Proceedings of the 9th international conference on Supercomputing
Informed multi-process prefetching and caching
SIGMETRICS '97 Proceedings of the 1997 ACM SIGMETRICS international conference on Measurement and modeling of computer systems
ELFSR0: object-oriented extensible file systems
PDIS '91 Proceedings of the first international conference on Parallel and distributed information systems
Statistical Language Learning
Fido: A Cache That Learns to Fetch
VLDB '91 Proceedings of the 17th International Conference on Very Large Data Bases
I/O Requirements of Scientific Applications: An Evolutionary View
HPDC '96 Proceedings of the 5th IEEE International Symposium on High Performance Distributed Computing
Theory, Volume 1, Queueing Systems
Theory, Volume 1, Queueing Systems
Predicting file system actions from prior events
ATEC '96 Proceedings of the 1996 annual conference on USENIX Annual Technical Conference
An analytical approach to file prefetching
ATEC '97 Proceedings of the annual conference on USENIX Annual Technical Conference
Profile-guided I/O partitioning
ICS '03 Proceedings of the 17th annual international conference on Supercomputing
An Adaptive Cache Coherence Protocol Specification for Parallel Input/Output Systems
IEEE Transactions on Parallel and Distributed Systems
Parallel Implementation of Back-Propagation Algorithm in Networks of Workstations
IEEE Transactions on Parallel and Distributed Systems
Mining block correlations to improve storage performance
ACM Transactions on Storage (TOS)
Source level transformations to improve I/O data partitioning
SNAPI '03 Proceedings of the international workshop on Storage network architecture and parallel I/Os
RACE: A Robust Adaptive Caching Strategy for Buffer Cache
IEEE Transactions on Computers
Hiding I/O latency with pre-execution prefetching for parallel applications
Proceedings of the 2008 ACM/IEEE conference on Supercomputing
Parallel I/O prefetching using MPI file caching and I/O signatures
Proceedings of the 2008 ACM/IEEE conference on Supercomputing
Cashing in on hints for better prefetching and caching in PVFS and MPI-IO
Proceedings of the 19th ACM International Symposium on High Performance Distributed Computing
Characterizing applications from power consumption: a case study for HPC benchmarks
ICT-GLOW'11 Proceedings of the First international conference on Information and communication on technology for the fight against global warming
Boosting Application-Specific Parallel I/O Optimization Using IOSIG
CCGRID '12 Proceedings of the 2012 12th IEEE/ACM International Symposium on Cluster, Cloud and Grid Computing (ccgrid 2012)
A New File-Specific Stripe Size Selection Method for Highly Concurrent Data Access
GRID '12 Proceedings of the 2012 ACM/IEEE 13th International Conference on Grid Computing
I/O acceleration with pattern detection
Proceedings of the 22nd international symposium on High-performance parallel and distributed computing
Hi-index | 0.02 |
Input/output performance on current parallel file systems is sensitive to a good match of application access patterns to file system capabilities. Automatic input/output access pattern classification can determine application access patterns at execution time, guiding adaptive file system policies. In this paper, we examine and compare two novel input/output access pattern classification methods based on learning algorithms. The first approach uses a feedforward neural network previously trained on access pattern benchmarks to generate qualitative classifications. The second approach uses hidden Markov models trained on access patterns from previous executions to create a probabilistic model of input/output accesses. In a parallel application, access patterns can be recognized at the level of each local thread or as the global interleaving of all application threads. Classification of patterns at both levels is important for parallel file system performance; we propose a method for forming global classifications from local classifications. We present results from parallel and sequential benchmarks and applications that demonstrate the viability of this approach.