Software-Extended Coherent Shared Memory: Performance and Cost
This paper evaluates the tradeoffs involved in the design of the software-extended memory system of Alewife, a multiprocessor architecture that implements coherentsharedmemory through a combination of hardware and software mechanisms. For each block of memory, Alewife implements between zero and fiv...
Other Authors: | |
---|---|
Format: | Text |
Language: | English |
Subjects: | |
Online Access: | http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.72.9545 http://www.cag.lcs.mit.edu/pub/papers/pdf/soft-ext-isca94.pdf |
id |
ftciteseerx:oai:CiteSeerX.psu:10.1.1.72.9545 |
---|---|
record_format |
openpolar |
spelling |
ftciteseerx:oai:CiteSeerX.psu:10.1.1.72.9545 2023-05-15T18:32:42+02:00 Software-Extended Coherent Shared Memory: Performance and Cost The Pennsylvania State University CiteSeerX Archives application/pdf http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.72.9545 http://www.cag.lcs.mit.edu/pub/papers/pdf/soft-ext-isca94.pdf en eng http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.72.9545 http://www.cag.lcs.mit.edu/pub/papers/pdf/soft-ext-isca94.pdf Metadata may be used without restrictions as long as the oai identifier remains attached to it. http://www.cag.lcs.mit.edu/pub/papers/pdf/soft-ext-isca94.pdf text ftciteseerx 2016-01-08T18:58:06Z This paper evaluates the tradeoffs involved in the design of the software-extended memory system of Alewife, a multiprocessor architecture that implements coherentsharedmemory through a combination of hardware and software mechanisms. For each block of memory, Alewife implements between zero and five coherence directory pointers in hardware and allows software to handle requests when the pointers are exhausted. The software includes a flexible coherence interface that facilitates protocol software implementation. This interface is indispensable for conducting experiments and has proven important for implementing enhancements to the basic system. Simulations of a number of applications running on a complete system (with up to 256 processors) demonstrate that the hybrid architecture with five pointers achieves between 71 % and 100% of full-map directory performance at a constant cost per processing element. Our experience in designing the software protocol interfaces and experiments with a variety of system configurations lead to a detailed understanding of the interaction of the hardware and software components of the system. The results show that a small amount of shared memory hardware provides adequate performance: One-pointer systems reach between 42 % and 100 % of full-map performance on our parallel benchmarks. A softwareonly directory architecture with no hardware pointers has lower performance but minimal cost. 1 Text The Pointers Unknown |
institution |
Open Polar |
collection |
Unknown |
op_collection_id |
ftciteseerx |
language |
English |
description |
This paper evaluates the tradeoffs involved in the design of the software-extended memory system of Alewife, a multiprocessor architecture that implements coherentsharedmemory through a combination of hardware and software mechanisms. For each block of memory, Alewife implements between zero and five coherence directory pointers in hardware and allows software to handle requests when the pointers are exhausted. The software includes a flexible coherence interface that facilitates protocol software implementation. This interface is indispensable for conducting experiments and has proven important for implementing enhancements to the basic system. Simulations of a number of applications running on a complete system (with up to 256 processors) demonstrate that the hybrid architecture with five pointers achieves between 71 % and 100% of full-map directory performance at a constant cost per processing element. Our experience in designing the software protocol interfaces and experiments with a variety of system configurations lead to a detailed understanding of the interaction of the hardware and software components of the system. The results show that a small amount of shared memory hardware provides adequate performance: One-pointer systems reach between 42 % and 100 % of full-map performance on our parallel benchmarks. A softwareonly directory architecture with no hardware pointers has lower performance but minimal cost. 1 |
author2 |
The Pennsylvania State University CiteSeerX Archives |
format |
Text |
title |
Software-Extended Coherent Shared Memory: Performance and Cost |
spellingShingle |
Software-Extended Coherent Shared Memory: Performance and Cost |
title_short |
Software-Extended Coherent Shared Memory: Performance and Cost |
title_full |
Software-Extended Coherent Shared Memory: Performance and Cost |
title_fullStr |
Software-Extended Coherent Shared Memory: Performance and Cost |
title_full_unstemmed |
Software-Extended Coherent Shared Memory: Performance and Cost |
title_sort |
software-extended coherent shared memory: performance and cost |
url |
http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.72.9545 http://www.cag.lcs.mit.edu/pub/papers/pdf/soft-ext-isca94.pdf |
genre |
The Pointers |
genre_facet |
The Pointers |
op_source |
http://www.cag.lcs.mit.edu/pub/papers/pdf/soft-ext-isca94.pdf |
op_relation |
http://citeseerx.ist.psu.edu/viewdoc/summary?doi=10.1.1.72.9545 http://www.cag.lcs.mit.edu/pub/papers/pdf/soft-ext-isca94.pdf |
op_rights |
Metadata may be used without restrictions as long as the oai identifier remains attached to it. |
_version_ |
1766216891737571328 |