Abstract. We investigate the use of runtime measurements to improve job scheduling on a parallel machine. Emphasis is on gang scheduling based strategies. With the information gathered at runtime, we define a task classification scheme based on fuzzy logic and Bayesian estimators. The resulting local task classification is used to provide better service to I/O bound and interactive jobs under gang scheduling. This is achieved through the use of idle times and also by controlling the spinning time of a task in the spin block mechanism depending on the node's workload. Simulation results show considerable improvements, in particular for I/O bound workloads, in both throughput and machine utilization for a gang scheduler using runtime information compared with gang schedulers for which this type of information is not available. 1
|
1486
|
Fuzzy sets
– Zadeh
- 1965
|
|
970
|
A bridging model for parallel computation
– Valiant
- 1997
|
|
240
|
Scheduling Techniques for Concurrent Systems
– Ousterhout
- 1982
|
|
184
|
Neural networks and fuzzy systems. A dynamical systems approach to machine intelligence. Englewood Cliffs, N.J
– Kosko
- 1992
|
|
106
|
Gang Scheduling Performance Benefits for Fine-Grained Synchronization
– Feitelson, Rudolph
- 1992
|
|
90
|
Improved utilization and responsiveness with gang scheduling
– Feitelson, Jette
- 1997
|
|
78
|
Distributed hierarchical control for parallel processing
– Feitelson, Rudolph
- 1990
|
|
74
|
Non-clairvoyant scheduling
– Motwani, Phillips, et al.
- 1994
|
|
70
|
et al, “PVM: Parallel Virtual Machine: A users’ guide and tutorial for networked parallel computing
– Geist
- 1994
|
|
59
|
Scheduling with Implicit Information in Distributed Systems
– Arpaci-Dusseau, Culler, et al.
- 1998
|
|
46
|
et al “LogP: Towards a Realistic Model of Parallel Computation
– Culler
- 1993
|
|
32
|
Coscheduling Based on Runtime Identification of Activity Working Sets
– Feitelson, Rudolph
- 1995
|
|
30
|
Performance Characteristics of Gang Scheduling In Multiprogrammed Environments
– Jette
- 1997
|
|
22
|
Lessons from Characterizing the Input/Output Behavior of Parallel Scientific Applications. Performance Evaluation: An
– Smirni, Reed
- 1998
|
|
17
|
Non-clairvoyant multiprocessor scheduling of jobs with changing execution characteristics
– Edmonds, Chinn, et al.
|
|
16
|
Mapping and scheduling in a shared parallel environment using distributed hierarchical control
– Feitelson, Rudolph
- 1990
|
|
14
|
Fuzziness vs. Probability
– Kosko
- 1990
|
|
13
|
Packing schemes for gang scheduling Job Scheduling Strategies for Parallel Processing
– Feitelson
- 1996
|
|
12
|
Overhead analysis of preemptive gang scheduling
– Hori, Tezuka, et al.
- 1998
|
|
12
|
Performance Evaluation of Gang Scheduling for Parallel and Distributed Multiprogramming
– Wang, Papaefthymiou, et al.
- 1997
|
|
8
|
Improving throughput and utilization in parallel machines through concurrent gang
– Silva, Scherson
- 2000
|
|
6
|
et al. Modeling of Workloads
– Jann
- 1997
|
|
6
|
A Lower Bound for Dynamic Scheduling of Data Parallel Programs
– Silva, Campos, et al.
- 1998
|
|
5
|
Solbalvarro et al. Dynamic Coscheduling on Workstation Clusters
– Patrick
- 1998
|
|
3
|
et al. Implementation of Gang Scheduling on Workstation Cluster. Job Scheduling Strategies for
– Hori
- 1996
|
|
3
|
Towards Flexibility and Scalability in Parallel Job Scheduling
– Silva, Scherson
- 1999
|
|
1
|
et al. A Practical Model of Parallel Computation
– Culler
- 1996
|
|
1
|
Gang Scheduling Performance Bene for Fine-Grain Synchronization
– Feitelson, Rudolph
- 1992
|
|
1
|
Coscheduling Based on Runtime Identi of Activity Working Sets
– Feitelson, Rudolph
- 1995
|
|
1
|
Decison Problems and Markov Chains
– Bayesian
- 1967
|
|
1
|
I/O Requirements of scienti c aplications: an evolutionary view
– Smirni, Aydt, et al.
- 1996
|
|
1
|
I/O Requirements of scientific aplications: an evolutionary view
– Smirni, Aydt, et al.
- 1996
|
|
