| M. Arunachalam, A. Choudhary, and B. Rullman. A prefetching prototype for the parallel file system on the Paragon. In Proceedings of the 1995. |
....in a natural manner and still achieve good performance even on out of core problem sizes. 5 Related Work Much related work has depended on the use of an explicit I O interface by the programmer. On the OS side, this work includes the automatic detection of file access patterns in the file system [1, 10, 11, 12, 14, 15, 17], as well as the use of access patterns supplied directly by the application using an I O type of interface [22, 26, 30, 4] For compilers it involves analysis to move explicit I O calls back and change them to asynchronous I O calls instead [25] While some of the OS policies developed may be ....
M. Arunachalam, A. Choudhary, and B. Rullman. A prefetching prototype for the parallel file system on the Paragon. In Proceedings of the 1995.
....in a natural manner and still achieve good performance even on out of core problem sizes. 5 Related Work Much related work has depended on the use of an explicit I O interface by the programmer. On the OS side, this work includes the automatic detection of file access patterns in the file system [1, 10, 11, 12, 14, 15, 17], as well as the use of access patterns supplied directly by the application using an I O type of interface [22, 26, 30, 4] For compilers it involves analysis to move explicit I O calls back and change them to asynchronous I O calls instead [25] While some of the OS policies developed may be ....
M. Arunachalam, A. Choudhary, and B. Rullman. A prefetching prototype for the parallel file system on the Paragon. In Proceedings of the
....knowledge to schedule future disk I O to reduce access latency, but this idea is not developed further. 7] presents results on prefetching into main memory by scheduling multiple prefetch commands for two disks on a SCSI bus, but the technique only works when the bus is not the bottleneck. [1] suggests that prefetching from disk into main memory cache by a parallel file system can be beneficial, but we note that the benefit of this approach is less clear when the bus is the bottleneck. 10] proposes and studies a scheduling algorithm for a video server that has multiple disks on each ....
Arunachalam, M., Choudhary, A., and Rullman, B. A prefetching prototype for the parallel file system on the Paragon. In Proceedings of ACM SIGMETRICS Conference on Measurement and Modeling of Computer Systems (May 1995), ACM Press, New York, NY, pp. 321--323. Extended Abstract.
....is only a hint to the system that may, or may not, be used in order to prefetch blocks from disk before the application actually needs them. In order to offer the user the possibility to implement their prefetching policies, some systems offer asynchronous reads as part of their interface [ACR95, GL91, HER 95] This kind of calls give the user all the needed flexibility to implement any kind of prefetching algorithm. 2.6.2 Prefetching via Data Compression A quite well studied prediction mechanism is the one used in data compression. Datacompression algorithms try to predict which ....
Arunachalam, M., Choudhary, A., and Rullan, B. A prefetching prototype for parallel file system on the Paragon. ACM Measurement and Modeling of Computer Systems (SIGMETRICS), 23(1):321--323, 1995.
....of disks. Kimbrel et al. show that the best behavior depends on the amount of contention of the disks, but that for reasonable data layouts on more than four disks contention is rarely a problem. Prefetching in file systems by automatically detecting file access patterns has been well studied [1, 11, 12, 13, 16, 17, 19]. Kroeger and Long look at using the compression technique known as prediction by partial match to detect access patterns and to decide what to prefetch [19] Griffioen and Appleton construct a probability graph based on Chapter 5. Related Work 68 prior file system accesses. Both approaches ....
M. Arunachalam, A. Choudhary, and B. Rullman. A prefetching prototype for the parallel file system on the Paragon. In Proceedings of the 1995 ACM Sigmetrics Conference on Measurement and Modeling of Computer Systems, pages 321--323, May 1995. Extended Abstract.
....and what to prefetch. The simplest algorithm is to load on every demand fetch the next adjacent block on disk: One Block Lookahead (OBL) Joseph, 1970] The OBLprediction engine was also used in databases [Smith, 1978b; Smith, 1978a; Smith, 1982; Smith, 1985] and similarly in parallel file systems [Arunachalam and Choudhary, 1995]. Two hardware solutions can be found in [Fu and Patel, 1991; Palacharla and Kessler, 1994] A strategy for a parallel merge sort algorithm was developed in [Pai and Varman, 1992; Pai et al. 1994; Wu et al. 1994] Simple strategies were already used in vector programs [Fu and Patel, 1991] ....
Arunachalam, M. and Choudhary, A. (1995). A Prefetching Prototype for the Parallel File System on the Paragon. In [SIGMETRICS, 1995], pages 321-- 323.
....in a natural manner and still achieve good performance even on out of core problem sizes. 5 Related Work Much related work has depended on the use of an explicit I O interface by the programmer. On the OS side, this work includes the automatic detection of file access patterns in the file system [1, 10, 11, 12, 14, 15, 17], as well as the use of access patterns supplied directly by the application using an I O type of interface [22, 26, 30, 4] For compilers it involves analysis to move explicit I O calls back and change them to asynchronous I O calls instead [25] While some of the OS policies developed may be ....
M. Arunachalam, A. Choudhary, and B. Rullman. A prefetching prototype for the parallel file system on the Paragon. In Proceedings of the 1995 ACM Sigmetrics Conference on Measurement and Modeling of Computer Systems, pages 321--323, May 1995. Extended Abstract.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC