Show simple item record

dc.creatorHolzrichter, Michael Warren
dc.date.accessioned2012-06-07T22:40:55Z
dc.date.available2012-06-07T22:40:55Z
dc.date.created1995
dc.date.issued1995
dc.identifier.urihttp://hdl.handle.net/1969.1/ETD-TAMU-1995-THESIS-H654
dc.descriptionDue to the character of the original source materials and the nature of batch digitization, quality control issues may be present in this document. Please report any quality issues you encounter to digital@library.tamu.edu, referencing the URI of the item.en
dc.descriptionIncludes bibliographical references.en
dc.descriptionIssued also on microfiche from Lange Micrographics.en
dc.description.abstractMemory architecture is an important component in a distributed shared-memory parallel computer. This thesis studies three shared-memory architectures-Non-Uniform Memory Access (NUMA) with full-mapped directories, Cache-Only Memory Architecture (COMA) with full-mapped directories, and COMA with directories based on a new design using binomial trees. The three architectures were implemented in the Proteus execution driven simulator. Proteus simulated the execution of three applications taken from the SPLASH-2 suite of benchmark parallel programs. Six sets of simulations were run. These simulations provided performance data for a range of values of important design parameters. The parameters studied were page size, block size, number of processors, memory controller speed, cache size and interconnection network topology. These simulations have two major benefits. First, they aid in choosing the best values for key design parameters. Second, these simulations facilitate the direct comparison of COMA vs. NUMA as well as the two directory designs. The simulations show that both COMA architectures generally perform better than NUMA. COMA proved to be less sensitive to suboptimum choices of primary cache and block sizes. In most cases the COMA with full-mapped directories performed a little better than with binomial trees. However, the binomial tree directories require significantly less hardware (eleven versus sixty-four bite per block for the machines simulated in this thesis).en
dc.format.mediumelectronicen
dc.format.mimetypeapplication/pdf
dc.language.isoen_US
dc.publisherTexas A&M University
dc.rightsThis thesis was part of a retrospective digitization project authorized by the Texas A&M University Libraries in 2008. Copyright remains vested with the author(s). It is the user's responsibility to secure permission from the copyright holder(s) for re-use of the work beyond the provision of Fair Use.en
dc.subjectelectrical engineering.en
dc.subjectMajor electrical engineering.en
dc.titlePerformance evaluation of NUMA and COMA distributed shared-memory multiprocessorsen
dc.typeThesisen
thesis.degree.disciplineelectrical engineeringen
thesis.degree.nameM.S.en
thesis.degree.levelMastersen
dc.type.genrethesisen
dc.type.materialtexten
dc.format.digitalOriginreformatted digitalen


Files in this item

Thumbnail

This item appears in the following Collection(s)

Show simple item record

This item and its contents are restricted. If this is your thesis or dissertation, you can make it open-access. This will allow all visitors to view the contents of the thesis.

Request Open Access