Experiences in building Cosy - an Operating System for Highly Parallel Computers

Butenuth, Roger; Burke, Wolfgang; Rose, C. De; Gilles, Sébastien; Weber, Reinhardt

doi:10.1016/s0927-5452(98)80061-0

Cited by 4 publications

(3 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Cosy microkernel operating system was designed and optimized for board-level multiprocessor and multicomputer platforms [55,56]. Cosy has a strong focus on providing the right platform abstractions to ease application development.…”

Section: Cosymentioning

confidence: 99%

A Safari Through the MPSoC Run-Time Management Jungle

Nollet

Verkest

Corporaal

2008

J Sign Process Syst

View full text Add to dashboard Cite

The multiprocessor SoC (MPSoC) revolution is fueled by the need to execute multiple advanced multimedia applications on a single embedded computing platform. At design-time, the applications that will run in parallel and their respective user requirements are unknown. Hence, a run-time manager (RTM) is needed to match all application needs with the available platform resources and services. Creating such a runtime manager requires two decisions. First, one needs to decide what functionality to implement. Second, one has to decide how to implement this functionality in order to meet boundary conditions like e.g. real-time performance. This paper is the first to detail a generic view on MPSoC run-time management functionality and its design space trade-offs. We substantiate the run-time components and the implementation tradeoffs with academic state-of-the-art solutions and a brief overview of some industrial multiprocessor run-time management examples. We show a clear trend towards more hardware acceleration, a limited distribution of management functionality over the platform and increasing support for adaptive multimedia applications.In addition, we briefly detail upcoming run-time management research issues.

show abstract

Section: Cosymentioning

confidence: 99%

A Safari Through the MPSoC Run-Time Management Jungle

Nollet

Verkest

Corporaal

2008

J Sign Process Syst

View full text Add to dashboard Cite

show abstract

“…The library is also portable except for a thin communication layer which has been adapted to MPI, PVM, COSY [43] and PARIX so far, so that it runs on most MIMD machines. The performance of the library at least equalizes previous less portable implementations.…”

Section: Portability Reusability Efficiencymentioning

confidence: 99%

Randomized Receiver Initiated Load-balancing Algorithms for Tree-shaped Computations

Sanders

2002

The Computer Journal

View full text Add to dashboard Cite

This paper considers generic load-balancing algorithms which efficiently parallelize a large class of applications based on traversing implicitly defined trees with irregular shape. First, a previous model is generalized yielding tree-shaped computations which cover the cost for communication and problem splitting, a measure of granularity and an easy to quantify parameter which limits irregularity. Then the random polling load-balancing algorithm is analyzed yielding upper bounds which match lower bounds for a large class of possible algorithms and machines. These bounds even hold for a fully asynchronous communication model which is important for practically efficient implementations. Then, with poll-and-shuffle, an asymptotically even more efficient algorithm is introduced. By using predominantly local communications, it increases the usable communication bandwidth on hypercubic networks and meshes by a logarithmic factor. These analytic results are complemented by practical refinements and implementation results which successfully apply a portable and reusable library on machines with up to 1024 processors.

show abstract

“…Many research systems like Amoeba [6], Locus [7], COSY [8] and Clouds [9] have established powerful primitives such as RPC, threads and group-communication that are required to develop such systems. Communication infrastructures like MPI [21] and OpenMP [26] have been developed for enabling parallelcomputing systems.…”

Section: Introductionmentioning

confidence: 99%

Distributed Resource Kernels: OS Support for End-To-End Resource Isolation

Lakshmanan

Rajkumar

2008

2008 IEEE Real-Time and Embedded Technology and Applications Symposium

View full text Add to dashboard Cite

The notion of resource reservation for obtaining real-time scheduling guarantees and enforcement of resource usage has gained strong support in recent years. However, much work on resource reservation has primarily focused on single-processor systems. In this paper, we propose the distributed resource kernel framework to deploy distributed real-time applications with end-to-end timing constraints, and to efficiently enforce and monitor their usage. Modern distributed real-time systems host multiple applications, where each application can span two or more processors. Timing bugs in one distributed application can affect the timing properties of other applications in the system. Our framework introduces the abstraction of a distributed resource container as an isolated virtual operating environment for a distributed real-time application. We have implemented this framework by extending our open-source singlenode Linux/RK platform [15]. A deployment and monitoring tool called dMon is also provided. We evaluate the framework's ability to provide timing guarantees by stress-testing the system using the Distributed Hartstone benchmarks. An audio processing pipeline is then used to illustrate the temporal isolation support provided by the Distributed RK framework. The distributed container abstraction can also be extended in the future to support security and fault-tolerance attributes.

show abstract

Experiences in building Cosy - an Operating System for Highly Parallel Computers

Cited by 4 publications

References 5 publications

A Safari Through the MPSoC Run-Time Management Jungle

A Safari Through the MPSoC Run-Time Management Jungle

Randomized Receiver Initiated Load-balancing Algorithms for Tree-shaped Computations

Distributed Resource Kernels: OS Support for End-To-End Resource Isolation

Contact Info

Product

Resources

About