Static communication scheduling for minimizing collisions in application specific parallel systems

Surma, David R.; Sha, Edwin Hsing-Mean

doi:10.1109/asap.1996.542819

Cited by 4 publications

(4 citation statements)

References 13 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Next, the same messages were input but in an order determined by a purely static scheduling algorithm called SAG. Details of this algorithm can be found in [9]. The schedule order and its results are also shown in Table 3.…”

Section: Definition 2 the Estimated Arrival Time Meat Is The Estimamentioning

confidence: 98%

“…Surma and Sha in their work addressed this problem both in a very restricted way and by studying a general case model of message traffic. However, they only considered static schedul- ing [9]. Surma, Tongsima and Sha showed in [10] how taking into account the communication costs can significantly improve the scheduling of cyclic data flow graphs [7].…”

Section: Introduc~onmentioning

confidence: 98%

“…Reducing the communication costs is the key issue in each case. By using a new model known as a Collision Graph [9], this paper develops a new scheduling framework able to deal with real-life problems.…”

Section: Introduc~onmentioning

confidence: 99%

See 2 more Smart Citations

Hybrid static-dynamic communication scheduling for parallel systems

Surma

Sha

1997

Proceedings of the 1997 ACM Symposium on Applied Computing - SAC '97

View full text Add to dashboard Cite

Parallel processing techniques have become widespread in applications requiring very high throughput or which have real-time deadlines. However, the potential gains from parallel processing can be diminished by the communication overhead inherent in these systems. With tightly-coupled architectures being used as the platform, an optimal way of scheduling the messages in a network is addressed. Static scheduling, while being able to utilize a priori information, is found to be lacking when this information is unavailable or inaccurate. Dynamic scheduling can adjust to changes within the network at run-time but suffers from not having any knowledge of the network traffic environment. To solve this problem, we introduce a hybrid scheduling technique which seeks to extract the best from each of these approaches. This hybrid scheduling technique incorporates a priority scheme derived from using the newly developed Collision Graph model. The determination of an optimal schedule is an NP-complete problem. Therefore, heuristics are used in the development of a priority mapping algorithm to deal with a general case model of message traffic. Experiments performed show a significant improvement over baseline approaches. INTRODUC~ONIn systems requiring high throughput or which have realtime deadlines, high-performance multi-processor designs are increasingly being used. While such systems can reduce the computation time, the communication overhead required can be substantial. One way to improve the total execution time is to hide this communication time by overlapping it with the computation time. The creation of a new scheduling "'Permission to make digital or hard copies of part or all of this work for personal or cl~sroom use is granted without lee provided that copies are not made or distributed lbr profit or commercial advantage and that copies bear this notice and the full citation on the first page. Copyrights lbr components of this work owned by others than ACM rot, st be honored. Abstracting with credit is permitted. To copy otherwise, to republish, to post on servers or to redistribute to lists, rcqt, ires prior specific permission and/or a lee." c 1997 ACM 0-897¢)1-850-9 97 0002 3.50 technique was required to achieve such a goal, since most of the existing scheduling methods do not consider the communication characteristics of the problem [4] and are unable to achieve an optimal schedule.In multi-processor systems, the performance is directly affected by the way tasks are allocated to the individual processing elements. Then, the communication costs stem from the underlying message passing. This research assumes that a suitable task allocation scheme, such as the one presented in [5], has been used and deals with the message transmissions. Therefore, the new scheduling technique is not a type of multi-processor scheduling, but rather deals strictly with determining an ordering of these transmissions. It does, howeven incorporate aspects of multi-processor scheduling in arriving at a novel hybrid static-dynamic sch...

show abstract

Section: Definition 2 the Estimated Arrival Time Meat Is The Estimamentioning

confidence: 98%

Section: Introduc~onmentioning

confidence: 98%

See 1 more Smart Citation

Hybrid static-dynamic communication scheduling for parallel systems

Surma

Sha

1997

Proceedings of the 1997 ACM Symposium on Applied Computing - SAC '97

View full text Add to dashboard Cite

show abstract

“…This paper uses a model presented by Surma and Sha [17], called a Collision Graph, to begin the research on the compile-time analysis and run-time scheduling of point-topoint message transmissions done to reduce the communication overhead. Work done in [17][18][19][20] addresses this problem in a restricted way by performing static scheduling using fixed routing. In [21,22] a dynamic scheduling technique was introduced but it also dealt only with fixed routing.…”

Section: Introductionmentioning

confidence: 99%

Communication Scheduling With Re-Routing Based on Static and Hybrid Techniques

Surma

Sha

Passos

2004

J CIRCUIT SYST COMP

View full text Add to dashboard Cite

In massively parallel systems, the performance gains are often significantly diminished by the inherent communication overhead. This overhead is caused by the required message passing resulting from the task allocation scheme. In this paper, techniques to reduce this communication overhead by both scheduling the communication and determining the routing that the messages should take within a tightly-coupled processor network are presented. Using the recently developed Collision Graph model, static scheduling algorithms are derived which work at compile-time to determine the ordering and routing of the individual message transmissions. Since a priori knowledge about the network traffic required by static scheduling may not be available or accurate, this work also considers dynamic scheduling. A novel hybrid technique is presented which operates in a dynamic environment yet uses known information obtained by analyzing the communication patterns. Experiments performed show significant improvement over baseline techniques.

show abstract

Simulink-Based Dynamic Priority Scheduling for Optimization in Network-on-Chip Communication

Pan

Shen

et al. 2008

2008 3rd International Conference on Innovative Computing Information and Control

View full text Add to dashboard Cite

Static communication scheduling for minimizing collisions in application specific parallel systems

Cited by 4 publications

References 13 publications

Hybrid static-dynamic communication scheduling for parallel systems

Hybrid static-dynamic communication scheduling for parallel systems

Communication Scheduling With Re-Routing Based on Static and Hybrid Techniques

Simulink-Based Dynamic Priority Scheduling for Optimization in Network-on-Chip Communication

Contact Info

Product

Resources

About