A survey of heterogeneous computing: concepts and systems

Ekmecic, I.; Tartalja, I.; Milutinović, Veljko

doi:10.1109/5.533958

Cited by 34 publications

(12 citation statements)

References 48 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…Heuristics in many cases prove capable of yielding near-optimal results in a signi®cantly shorter time than full search. Heterogeneous Optimal Selection Theory [4], for example, proposes a hierarchical but heuristical approach.…”

Section: Relation To Other Workmentioning

confidence: 99%

Fuzzy configuration of matching runtime implementation strategies

Sodan¹,

Torra²

2001

Soft Computing

View full text Add to dashboard Cite

With applications currently growing in complexity and range, increasing numbers of con®guration problems are arising in compilers. Already many software systems offer multiple specialized implementation strategies and substrategies, differing in terms of applicability and/or cost, depending on the application context. Con®gurations then have to be created from the different strategies available in accordance with the application characteristics, the global optimization objective, and potential constraints on the strategies' combinability. In many cases, this results in a combinatorial, i.e., discrete, optimization problem. Proper solutions for automating the con®guration while limiting the complexity of the solution search are still being sought. We address here the ®eld of parallel/distributed processing and the con®gura-tion of runtime implementation strategies, such as for communication or dynamic load balancing. We present a rule-based approach, integrating fuzzy methodologies for the classi®cation of application characteristics and for gradual selection preference in rules. In this way we exploit available knowledge about the correlation of the problem and solution space, and apply soft computing methods to obtain an approximate, rather than perfect, solution approach, thus helping to limit the con®guration complexity. Our approach extends standard fuzzy inference by a multistage organization, and ± with proper organization of rules, characteristics and strategies ± performs hierarchical fuzzy inference. The approach is demonstrated on concrete con®guration examples in parallel compilers.

show abstract

Section: Relation To Other Workmentioning

confidence: 99%

Fuzzy configuration of matching runtime implementation strategies

Sodan¹,

Torra²

2001

Soft Computing

View full text Add to dashboard Cite

show abstract

“…Dentro de la computación paralela se encuentra la computación heterogénea [17,18], que se refiere al uso de sistemas o entornos de computación que están compuestas por unidades de procesamiento de diferente naturaleza. Algunos ejemplos clásicos son los convencionales sistemas de memoria compartida o distribuida que contienen multi-core CPUs junto con dispositivos many-core como las GPUs.…”

Section: Examples Of a Graph Withunclassified

“…Heterogeneous computing [17,18] refers to computing environments or systems that are built with computational devices of different nature. Some classic examples are conventional shared-or distributed-memory multi-processors containing multi-core CPUs, together with GPU devices.…”

Section: Heterogeneous Computingmentioning

confidence: 99%

“…The normal CPU cores also enter into the ask-for-tasks working loop but executing the code of the pluginCPU (see Sect. 6.3.2) (lines [16][17][18]. Lines 14 and 17 show the code responsible for obtaining the following task to be executed by invoking the pluginSLAVE macro.…”

Section: Cluster In-node Synchronization (2nd Layer): Tuccompi_parallelmentioning

confidence: 99%

“…If the requesting slave is an old GPU (pre-Fermi) or a CPU core, the master only sends a single task to the slave (lines 12-15), thus, the task counter is simply incremented. When all tasks have been scheduled and carried out, the master sends the termination signal to the rest of active slaves when they request more tasks (lines [17][18][19][20][21].…”

Section: Scheduling Plug-insmentioning

confidence: 99%

See 2 more Smart Citations

Parallel approaches to shortest-path problems for multilevel heterogeneous computing

Arranz¹

View full text Add to dashboard Cite

Modelo paralelo abstracto, APSP, tuning automático de kernels, Boost Graph Library, configuración de la Cache L1, ejecución concurrente de kernels, CUDA, Dijkstra, GPGPU, sistemas heterogéneos, framework HPC, modelo de caracterización de kernels, proceso de caracterización de kernels, balanceo de carga, MPI, comparativa de plataformas de NVIDIA, OpenMP, técnicas de optimización, algoritmos paralelos, SSSP, geometría del bloque de hilos.

show abstract

Heterogeneous Distributed Computing

Maheswaran

Braun

Siegel

1999

Wiley Encyclopedia of Electrical and Electronics Engineering

View full text Add to dashboard Cite

One of the biggest challenges with high-performance computing is that as machine architectures become more advanced to obtain increased peak performance, only a small fraction of this performance is achieved on many real application sets. This is because a typical application may have various subtasks with different architectural requirements. When such an application is executed on a given machine, the machine spends most of its time executing subtasks for which they are unsuited. With the recent advances in high-speed digital communications, it has become possible to use collections of different high-performance machines in concert to solve computationally intensive application tasks. This article describes the issues involved with using such a heterogeneous computing (HC) suite of machines to solve application tasks.A hypothetical example application that has various subtasks, which are best suited, for different machine architectures is shown in Figure 1 (based on (FrS93)). The example application executes for 100 time units on a baseline serial machine. The application consists of four subtasks: the first is best suited to execute on an SIMD (synchronous) parallel machine, the second is best suited for a distributed-memory MIMD (asynchronous) parallel machine, the third is best suited for a shared-memory MIMD machine, and the fourth is best suited to execute on a vector (pipelined) machine.Executing the whole application on an SIMD machine may improve the execution time of the SIMD subtask from 25 to 0.01 time units, and the other subtasks to varying extents. The overall execution time improvement may only be about a factor of five because other subtasks may not be well suited for an SIMD machine. Using four different machines that match the computational requirements for each of the individual subtasks can result in an overall execution time that is

show abstract

A survey of heterogeneous computing: concepts and systems

Cited by 34 publications

References 48 publications

Fuzzy configuration of matching runtime implementation strategies

Fuzzy configuration of matching runtime implementation strategies

Parallel approaches to shortest-path problems for multilevel heterogeneous computing

Heterogeneous Distributed Computing

Contact Info

Product

Resources

About