Showing papers on "Task (computing) published in 1982"

A Task Allocation Model for Distributed Computing Systems

Abstract: This paper presents a task allocation model that allocates application tasks among processors in distributed computing systems satisfying: 1) minimum interprocessor communication cost, 2) balanced utilization of each processor, and 3) all engineering application requirements.

Method of processing an operating system in a multi-processor system

Abstract: An operating system (OS) is divided into units (OS processes) which are concurrently executable. When the processors concurrently request the execution of one OS process, data indicating that the requested OS process is in a ready state is loaded into a memory. The other processors repeatedly make an access to the data during the course of the execution of the user's task to check to see whether the ready-state OS process is present or not. A processor when detecting the ready-state OS process ceases the execution of the user's program and executes that OS process.

Software Structures for Ultraparallel Computing

Abstract: In this thesis we implement several basic parallel processing primitives by using a "replace-add" operation, which can supersede the standard "test and set", and which appears to be a universal primitive for efficiently coordinating large numbers of independently acting sequential processors. The replace-add is essentially an indivisible add-to-memory operation although concurrent replace-adds can all be processed in the same one cycle. In particular, we use the replace-add to develop routines for concurrent access to a queue and show how they can be used to devise many highly parallel algorithms as well as a distributed, concurrent task scheduler. The "paracomputer" forms our underlying theoretical model of parallel computation although we also consider a realistic architecture approximating this model. We justify our use of the replace-add operation by presenting a hardware implementation that permits multiple replace-adds to be processed nearly as efficiently as loads and stores. Moreover, the crucial special case of concurrent replace-adds updating the same variable is handled particularly well: If every PE simultaneously addresses a replace-add at the same variable, all these requests are satisfied in the time required to process just one request.

Minimizing Expected Makespan in Stochastic Open Shops.

Abstract: : Suppose that two machines are available to process n tasks. Each task has to be processed on both machines, the order in which this happens is immaterial. Task j has to be processed on machine 1 (2) for random time X sub j (Y sub j) with distribution F sub j (G sub j). This kind of model is usually called an Open Shop. The time that it takes to process all tasks is normally called the makespan. Every time a machine finishes processing a task the decision-maker has to decide which task to process next. Assuming that X sub j and Y sub j have the same exponential distribution we show that the optimal policy instructs the decision-maker, whenever a machine is freed, to start processing the task with longest expected processing time among the tasks still to be processed on both machines. If all tasks have been processed at least once, it does not matter what the decision-maker does, as long as he keeps the machine busy. We then consider the case of n identical tasks and two machines with different speeds. The time it takes machine 1 (2) to process a task has distribution F(G). Both distributions F and G are assumed to be New Better than Used (NBU) and we show that the decision-maker stochastically minimizes the makespan when he always gives priority to those tasks which have not yet received processing on either machine. (Author)

Object and task reclamation in distributed applicative processing systems

Abstract: Functional programming has received increased attention in recent years, for reasons related to more effective programming and the potential for highly-parallel implementations. This dissertation is concerned with the general problem of automatic storage reclamation in functional programming systems utilizing distributed heaps. In such "distributed applicative processing systems," the evaluation of a functional program is accomplished by a highly-parallel graph-reduction process, effectively distributed among many processing elements having only local store. This evaluation process creates the potential for unique forms of "garbage," whose detection may be crucial to effective performance. In particular, there may be "dormant" portions of the computation graph (equivalent to deadlock), in addition to "irrelevant tasks" (active processes which are determined to be no longer relevant to the computation). There are three primary contributions of this research. First, the model of distributed applicative processing systems, together with the characterization of irrelevant tasks and dormant subgraphs, is useful in generalizing the standard storage reclamation problem to highly parallel, distributed systems. The deficiencies of existing storage reclamation schemes in such an environment are discussed in detail. Second, the marking-tree collector is presented as the first effectively distributed real-time garbage collector of the mark-sweep variety. This scheme does not require centralized data or control (other than a logical rendezvous between phases of the collector) and operates in parallel with the main computation. A copying version of the scheme is also presented, with improved performance. Lastly, the technique of call-graph reclamation is presented, which demonstrates the feasibility of doing efficient storage reclamation at the block level, by making "dynamic" reclamation decisions based on the types of intermediate results in a computation. All of the schemes presented have the added benefit of being able to detect, and subsequently delete, irrelevant tasks.

Automatic/Control Processing Concepts and Their Implications for the Training of Skills.

Abstract: : Recent research on automatic and controlled processing is reviewed with emphasis on the implications to applied training programs and personnel selection. The research assumes that humans perform mental operations through the interaction between slow, effortful, easy to modify controlled processing and comparatively fast, difficult to train automatic processing. Major conclusions from the work include: performance improves as a function of consistent correct executions; automatic processes can develop with few training trials; massing or distributing practice has little effect; reducing control processing resources slows learning; automatic processes can become context dependent; training operators to respond to classes of events promotes generalization of automatic processing; dual task measures are important assessment tools in assessing skill competency; and motivational characteristics of training programs greatly influence training success rates. Automatic processes influence performance in complex category search tasks, inconsistent tasks with consistent components, task which are consistent within a context, visual spatial temporal pattern tasks, and conjunction of features tasks. Automatic processing is very reliable and can perform complex functions, such as category search, with no measurable increase in workload.

The Array Size Function in Simple Visual Search Tasks: A Comparison Between "'go-no go" and "'yes-no" Tasks Under Conditions of High and Low Target-noise Similarity

Abstract: In research on visual search within a single eye-fixation a number of different tasks are used and referred to interchangeably. Research with other types of tasks suggests that there are possibly important differences between these tasks. In the present study, two types of search tasks were compared under conditions as equal as possible: the ‘go-no go’ task and the ‘yes-no’ task. Conditions of low and high target-noise similarity were used. The results obtained showed: a) a steeper slope of the array size function in ‘yes-no’ tasks than in ‘go-no go’ tasks on the first day of practice but not on the second: b) a higher intercept value of the same function for ‘yes-no’ tasks than for ‘go-no go’ tasks; and c) a greater proportion of errors with ‘yes-no’ tasks than with comparable ‘go-no go’ tasks. A tentative model, describing the main features of the results obtained, is briefly sketched.

Message transmission and reception system between processor

Abstract: PURPOSE:To reduce an overhead by sending a request signal while placing a task at a transmission destination in an execution waiting state by a transmission-side processor in case of message transmission and reception between processors, and by performing diapatch at a reception side according to priority CONSTITUTION:When messages are transmitted from 10 subsystems 4a and 4b to central processor subsystems 3a and 4b, dispatch FFs 8a-8d in respective processors 3a and 3b, and 4a and 4b are set, and a task with top priority is dispatch in the queue of the transmission-destination task execution of the subsystems 4a and 4b to which those processors belong to change a message queue into an execution queue Then a request to dispatch is sent out to the reception side, where the processor dispatches the task which corresponds to priority

The Use of Machine Aids in Dynamic Multi-Task Environments: A Comparison of an Optimal Model to Human Behavior.

Abstract: : Due to developments in the fields of computer and microprocessor technology, more and more tasks that previously only could be done by humans can now be assigned to machines. This study explores cases where such smart machines act as aids to human operators in dynamic multi-task systems. The operator must act as a supervisor and assign these machine aids to tasks he discovers. (Author)

Controlling system of computer system

Abstract: PURPOSE:To reduce the overhead, by controlling applications, a high-speed response is required, by the first operating system and controlling applications, where a severe response is required, by the second operating system. CONSTITUTION:The first operating system OS (nucleus OS) 41 controls the second OS (secondary OS) task 44 in the same relation as other user task groups 42. Secondary task groups 44-2 are operated under the control of a secondary OS 44-1 by multiprogramming. Three principal entries, namely, an initial entry 44-4, a trap entry 44-3, and an interruption entry 44-5 are provided as entries to the OS task 44, and thus, the secondary OS 44-1 has a function sufficient for controlling task groups 44-2. Consequently, the nucleus OS has not the extra increase of overhead and load due to the existence of the secondary OS.

Operating system for terminal equipment

Abstract: PURPOSE:To make the assignment and the revervation of input/output devices efficient, by generating queue structures for every level of respective tasks in accordance with their state and ranking tasks in these queue and changing this ranking. CONSTITUTION:Lower-level tasks 2 and 2' consisting of tasks for interpreting user programs, middle-level tasks 3 and 3' which control input/output devices 5 and 5', and a high-level task 1 which controls the state of these low-level and middle-level tasks are controlled by an operating system OS 4. Respective tasks are registered in an execution queue (E.Q) or ready queue (R.Q) for every level, and ranks of execution waiting tasks which are ranked and registered in the R.Q in the FIFO system are changed by a rank changing means. Commands issued to waiting middle-level tasks are registered in a command execution list, and commands issued to executing or execution waiting middle-tasks are registered in a command start list. Consequently, input/output devices are used efficiently.

Monitor device for process state

Abstract: PURPOSE:To realize a high-speed sampling, by providing a direct task, which is not controlled by the OS, and a timer which prescribes the sampling period and outputting a task start signal from the input/output device for the input of a process state. CONSTITUTION:An OS start level interrupt signal 14 which is outputted from a process 1 through a process input/output device or is outputted from an operation panel 17 starts an OS 15 first, and the OS 15 starts general tasks 19 under its control. Meanwhile, a direct task start level interrupt signal 13 which is outputted from the process 1 through the process input/output device drives hardware in a level higher than that for the OS 15 to start a high-speed sampling task (HSS) 18, which is not under the control of the OS 15, without intervention of the OS 15. In figure, 18 and 20 indicate the sampling task and an information processing task respectively.

Experimental evaluation of the concept of supevisory manipulation

Abstract: A computer-controlled teleoperator system which is based on task-referenced sensor-aided control has been developed to study supervisory manipulation. This system, called SUPERMAN, is capable of performing complicated tasks in real-time by utilizing the operator for high-level functions related to the unpredictable portions of a task, while the subordinate machine performs the more well-defined subtasks under human supervison. To determine whether supervisory control schemes such as these offer any advantage over manual control under real-time conditions, a number of experiments involving both simple and complicated tasks were performed. Six representative tasks were chosen for the study: (1) obtaining a tool from a rack, (2) returning the tool to the rack, (3) removing a nut, (4) placing samples in a storage bin, (5) opening and closing a valve, and (6) digging with a shovel. The experiments were performed under simulated conditions using four forms of manual control (i.e., switch rate, joystick rate, master-slave position control, and master-slave with force feedback), as well as supervisory control. Through these experiments the effectiveness and quality of control were evaluated on the basis of the time required to complete each portion of the task and the type and number of errors which occurred.

Data logging system

Abstract: A data logging system in which there are two data buses 3,3a each having one of a pair of units coupled to it. One unit of the pair is committed to a particular task and the other unit of the pair is non-committed and works asynchronously with the committed unit. For each committed unit there is a verification means which in the event of malfunctioning of the committed unit reassigns the tasks performed by that committed unit to the corresponding non-committed unit. The units may be tape recorders 5,5a, data stores 6,6a, computers 7,7a and displays 8,8a.