1. © ANALYSIS OF MACRO-DATA-FLOW DYNAMIC SCHEDULING ON NONUNIFORM MEMORY ACCESS ARCHITECTURES ALMOUHAMED, M IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC, IEEE TRANSACTIONS ON PARALLEL AND DISTRIBUTED SYSTEMS; pp: 875-888; Vol: 4 King Fahd University of Petroleum & Minerals http://www.kfupm.edu.sa Summary This paper studies dynamic scheduling of computational tasks with communication costs using nonuniform memory access architecture.The computing model assumes that data transfer can be partitioned into parallel and sequential parts with respect to the task execution. A scheduling heuristic, called least-communication (LC), together with a two-level scheduler is proposed in an attempt to minimize the finish time. The LC selects the task that removes the largest amount of remaining data transfer, if no such tasks are available the task that has been ready to run at the earliest is selected first. Analysis shows that scheduling n tasks on m processors gives a schedule length omega(lc) that satisfies omega(lc) less-than-or-equal-to (2 - 1/m)omega(opt) + DELTAtSIGMAC(S)(T(i), T(i+1)) + (n/m + (1 - 1/m)L)(h(lc) - h(opt)), where omega(opt) is the length of an optimum solution for scheduling with no communication costs, SIGMAC(S) is the sum of sequential data transfer for all the arcs (T(i), T(i+1)) along a chain of L tasks whose computation and data transfer times cover all the idle times in the schedule, DELTAt is the time to transfer one operand to memory, and h(opt) and h(lc) are the task scheduling overheads in case of optimum and LC scheduling, respectively.The time complexity of LC is O(n2). Testing the finish time of LC and FCFS shows that LC is useful for tasks having moderate granularity and whose computation and communication requirements vary widely for different data sets. Copyright: King Fahd University of Petroleum & Minerals; http://www.kfupm.edu.sa

2. 1. 2. 3. 4. 5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. © References: 1985, BUTTERFLY PARALLEL P 1988, BUTTERFLY GP1000 ADAM TL, 1974, COMMUN ACM, V17, P685 ALMOUHAMED M, 1990, MICROPROCESS MICROSY, V14, P276 ALMOUHAMED MA, 1990, IEEE T SOFTWARE ENG, V16, P1390 CHEN CL, 1988, IEEE T SYST MAN CYB, V18, P729 COFFMAN EG, 1973, OPERATING SYSTEMS TH CVETANOVIC Z, 1987, IEEE T COMPUT, V36, P421 DENNIS JB, 1980, J COMPUT, V13 ELREWINI H, 1990, J PARALLEL DISTR COM, V9, P138 FERNANDEZ EB, 1973, IEEE T COMPUT, V22, P745 GAREY MR, 1978, OPER RES, V26, P3 GONZALEZ MJ, 1977, ACM COMPUT SURV, V9, P173 GOTTLIEB A, 1983, IEEE T COMPUT, V32, P175 HWANG JJ, 1989, SIAM J COMPUT, V18, P244 KASAHARA H, 1984, IEEE T COMPUT, V33, P1023 LEE EA, 1990, J PARALLEL DISTRIBUT, P333 LO VM, 1984, 4TH P INTC DISTR COM LO VM, 1988, IEEE T COMPUT, V37, P1384 MCGREARY C, 1989, COMMUN ACM, V32, P1073 OFFMAN EG, 1976, COMPUTER JOB SHOP SC PASE DM, 1989, THESIS OREGON GRADUA PLAS A, 1976, 1976 P INT C PAR PRO, P293 RAMAMOORTHY CV, 1972, IEEE T COMPUT, V21, P137 SHIRAZI B, 1990, J PARALLEL DISTR COM, V10, P222 SMITH BJ, 1981, REAL TIME SIGNAL PRO, V298, P241 For pre-prints please write to: abstracts@kfupm.edu.sa Copyright: King Fahd University of Petroleum & Minerals; http://www.kfupm.edu.sa