IEEE Transactions on Parallel and Distributed Systems
Future Generation Computer Systems - Special issue on metacomputing
Static scheduling algorithms for allocating directed task graphs to multiprocessors
ACM Computing Surveys (CSUR)
A comparison of list schedules for parallel processing systems
Communications of the ACM
Application-level scheduling on distributed heterogeneous networks
Supercomputing '96 Proceedings of the 1996 ACM/IEEE conference on Supercomputing
Multiparadigm communications in Java for grid computing
Communications of the ACM
Automatic Coarse Grain Task Parallel Processing on SMP Using OpenMP
LCPC '00 Proceedings of the 13th International Workshop on Languages and Compilers for Parallel Computing-Revised Papers
The Globus Project: A Status Report
HCW '98 Proceedings of the Seventh Heterogeneous Computing Workshop
Scheduling with Advanced Reservations
IPDPS '00 Proceedings of the 14th International Symposium on Parallel and Distributed Processing
Practical Multiprocessor Scheduling Algorithms for Efficient Parallel Processing
IEEE Transactions on Computers
On the influence of network characteristics on application performance in the grid environment
ICN'05 Proceedings of the 4th international conference on Networking - Volume Part I
| Hi-index | 0.00 |
In the Grid applications, it is very likely that object oriented languages, like Java and Ruby, will be employed as well as large-scale semi-structured data like XML. However, their inherent dynamic memory management has to suspend the execution of all the tasks running on processors when it is invoked. This will adversely affect the Grid computing severely unless the task scheduling system can avoid it with some special mechanism.In this paper, we propose a new task scheduling method, referenced to as CP/MM, which can efficiently schedule tasks for applications requiring memory management. The underlying concept is to consider the cost due to memory management when the task scheduling system allocates ready coarse grain tasks, or macro-tasks, to processors. We have developed three task scheduling modules including an implementation of CP/MM into a task scheduling systemis implemented on Java RMI (Remote Method Invocation) communication infrastructure. Moreover, we evaluate the fundamental performance of CP/MM in two ways. The first is the performance evaluation of CP/MM which is applied to a small but practical test application program on a PC cluster. The second is the performance evaluation for test programs which have many tasks with complicated dependency relations on the test bed system consisting of computers on our two campuses located at a distance of approximately 32 km. These experimental results show that CP/MM can successfully prevent high priority macro-tasksfrom being affected by the garbage collection arising from the memory management, so that CP/MM can efficiently schedule distributed programs whose critical paths are relatively long.