Damian Kopanski

Graphical Design of Parallel Programs with Control Based on Global Application States Using an Extended P-Grade System

An extension of the graphical parallel program design system P-GRADE towards specification of program execution control based on global application state monitoring is presented. De-coupled structured specifications of computational and control elements of parallel programs are assumed. Special synchronizer processes collect process state messages supplied with time interval timestamps and construct strongly consistent application states. Control predicates are evaluated on these states by synchronizers. As a result, control signals can be sent to application processes to stimulate desired reactions to the predicates. The signals can cause asynchronous computation activation or cancellation. Implementation of a parallel program of Traveling Salesman Problem (TSP) solved by branch-and-bound (B&B) method is described to illustrate properties of the new system.

Program Design Environment for Multicore Processor Systems with Program Execution Controlled by Global States Monitoring

A new distributed program graphical design environment is described in the paper. It is oriented towards designing program execution control based on a built-in system infrastructure which enables easy global application states monitoring in systems based on multicore processors. Two aspects of global application control design are covered. First is the global control flow in programs at the level of processes and threads. The second is the asynchronous control of internal process and thread behavior. The proposed control infrastructure is based on structural program elements called synchronizers organized at the process and thread levels to collect state information, evaluate control predicates on global states and send signals to application program threads and processes to stimulate global control actions. The paper presents principles of the application program graphical design and programming methods to implement global control at the level of threads.

Global predicate monitoring applied for control of parallel irregular computations

Many computational problems have irregular data/control characteristics, which make programs difficult to be efficiently implemented in parallel systems. Due to irregular character of code or data, even division of work between processors at application startup is frequently impossible. Runtime optimization is possible, but it requires a constant exchange of control information and/or data during runtime is required. A novel parallel application control method is proposed in the paper. It is based on application global state monitoring for runtime irregular application control. The method provides a ready-to-use control infrastructure, which can be conveniently applied by a programmer. Both suitability and efficiency of the proposed control method are discussed in the paper based on two selected numerical applications: adaptive integration and branch and bound search. The presented experimental results were obtained with PS-GRADE graphical parallel design system, which embeds the proposed control method. The results confirm the efficiency of control based on global predicates in irregular computations

Implementing control in parallel programs by synchronization-driven activation and cancellation

Implementation of an advanced synchronization environment for multi process parallel applications has been proposed in the paper. Synchronization-driven asynchronous activation and cancellation in parallel program execution is first discussed. Execution activation and cancellation result from synchronization signals generated by a system of distributed synchronization agents that monitor and examine states of parallel applications. A proposal is given to integrate such a synchronization environment with the existing GRADE graphical parallel program design tool. The implementation design issues are assessed by preliminary performance tests in different executive environments.

Adding advanced synchronization to processes in GRADE

New synchronization mechanisms using asynchronous computation activation and cancellation, based on state monitoring of a parallel application, are presented. The paper proposes how to integrate the mechanisms with the GRADE graphical parallel program design environment. Necessary enhancements to the GUI along with the semantics and possible applications are presented. Efficient implementation methods for the proposed synchronization are discussed.

Parallel Irregular Computations Control Based on Global Predicate Monitoring

Global state monitoring can be applied in parallel applications in cluster environments to efficiently control execution of constituent processes. We present an application control method based on global predicate monitoring, used in a parallel system with partially synchronized computer clocks. Thanks to the clock synchronization, the global state construction can be done without big overhead. The proposed control method has been implemented in an enhanced P-GRADE graphical parallel program design system (PS-GRADE). As many irregular computational problems require sophisticated and unpredictable control actions, their implementations can benefit from the proposed control method. The traveling salesman problem and adaptive integration - two irregular applications - have been implemented in PS-GRADE. Obtained results have shown that the control by global predicate monitoring can lead to better parallel application performance and provides a convenient infrastructure for a programmer

Co-Ordination of Parallel GRID Applications using Synchronizers

Recently, the graphical parallel programming environment P-GRADE based on message passing has been extended by advanced synchronization and control mechanisms (synchronizers) based on predicates computed on consistent application global states (PS-GRADE). In the new P-GRADE Workflow system, the application workflow has been introduced to enable designing control flow between otherwise independent applications. In this paper, we propose that program execution control by synchronizers can be extended and used to coordinate parallel applications executed on GRID in the frame of a PS-GRADE tool. We introduce explicit inter-application, or GRID-level, synchronizers. They are connected to application level synchronizers, which are able to send state reports and receive control signals over GRID communication network. Open Grid Services Infrastructure, implemented using Globus toolkit, is proposed to be used for these purposes.

Dynamic Load Balancing Based on Applications Global States Monitoring

The paper presents how to use a special novel distributed program design framework with evolved global control mechanisms to assure processor load balancing during execution of applications. The new framework supports a programmer with an API and GUI for automated graphical design of program execution control based on global application states monitoring. The framework provides high-level distributed control primitives at process level and a special control infrastructure for global asynchronous execution control at thread level. Both kinds of control assume observations of current multicore processor performance and communication throughput enabled in the executive distributed system. Methods for designing processor load balancing control based on a system of program and system properties metrics and computational data migration between application executive processes is presented and assessed by experiments with execution of graph representations of distributed programs.

Distributed Program Execution Control Based on Application Global States Monitoring in PEGASUS DA Framework

This paper presents control implementation methods for an original distributed program design framework PEGASUS DA (Program Execution Governed by Asynchronous SUpervision of States in Distributed Applications) which provides automated design of distributed program execution control based on program global states monitoring. The framework includes a built in support for handling local and global application states as well as automatic construction and use of strongly consistent application global states for program execution control. In particular, the paper presents methods used to implement distributed program control inside the PEGASUS DA framework run on clusters of contemporary multicore processors based on multithreading. The program design method is illustrated on a distributed multithreaded application executed with load balancing in a multicore system.

Dynamic Workflow Control with Global States Monitoring

This paper shows how to organize dynamic workflow execution control based on predicates computed on internal global states of activities. Such paradigm has been implemented using special control processes called synchronizers. The proposed ideas extend the ordinary task and application workflow schemes by providing dynamic functional features. The proposed control has been provided in a graphical parallel program design environment, which includes a flexible infrastructure for designing advanced cluster or Grid level distributed programs. The paper presents a description of the proposed Grid programming environment and explains involved dynamic workflow implementation features.