The high performance solution of irregular problems

Dew, Peter M.; Nash, Jonathan M.

doi:10.1109/mppm.1997.715965

Cited by 1 publication

(6 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The Lock is not directly visible to the application in this case, since it is used to support Accum and PriQueue. The implementation issues for the Lock and Accumulator are described in [3]. The PriQueue implementation is described in more detail in Section 3.…”

Section: The Travelling Salesman Problemmentioning

confidence: 99%

“…One point to note is that the PriQueue stores the tasks within a highly concurrent shared data structure, and therefore requires a high performance network for good performance. An alternative implementation of the PriQueue, which is more amenable to networks where communication is more costly, is outlined in [3]. …”

Section: The Travelling Salesman Problemmentioning

confidence: 99%

“…Looking back to Figure 2(b), the local data segments are accessed cyclically by the processors, which can be supported by a SHMEM concurrent atomic increment operation, Inc (taken modulus p to obtain the required PriQueue data segment). Each data segment contains a number of locks and local flags, which are used to control the access to the data segments [3]. The implementation of these locks, through a lower level Lock SADT [3], can use the SHMEM atomic Swap operation, in which a new value is exchanged with the current contents of a shared word on a given processor.…”

Section: Performance Of the Shmem Librarymentioning

confidence: 99%

“…Each data segment contains a number of locks and local flags, which are used to control the access to the data segments [3]. The implementation of these locks, through a lower level Lock SADT [3], can use the SHMEM atomic Swap operation, in which a new value is exchanged with the current contents of a shared word on a given processor. This operation allows for the construction of a scalable linked list [8].…”

Section: Performance Of the Shmem Librarymentioning

confidence: 99%

“…This paper describes work on supporting irregular problems with scalable high performance, while preserving the BSP-style cost model. The key idea has been to develop scalable Shared Abstract Data Types (SADTs) which support dynamic sharing patterns [10,3]. An SADT is an extension of a serial ADT which allows the concurrent invocation of its methods.…”

Section: Introductionmentioning

confidence: 99%

See 4 more Smart Citations

Using SADTs to support irregular computational problems

Nash

Dew

Berzins

Proceedings Fourth International Symposium on Parallel Architectures, Algorithms, and Networks (I-Span'99)

Self Cite

View full text Add to dashboard Cite

There are well defined methods for supporting regular problems with scalable performance, typified by the HPF language and the BSP model. Less well understood is the solution of more irregular problems, supporting complex shared data structures and task dependencies, and typically requiring dynamic load balancing to sustain high performance. It is demonstrated how the use of Shared Abstract Data Types (SADTs), together with an extended BSP cost model, can support irregular problems in a structured manner. An SADT is an extension of a serial ADT which allows the concurrent invocation of its methods. A number of SADTs are used to implement a solution of the travelling salesman problem on the Cray T3D machine, and a description of the restructuring of a parallel CFD code using SADTs is provided, with initial results given for the Cray T3E.

show abstract

Section: The Travelling Salesman Problemmentioning

confidence: 99%

Section: The Travelling Salesman Problemmentioning

confidence: 99%