2012
DOI: 10.1145/2086696.2086704
|View full text |Cite
|
Sign up to set email alerts
|

Dapsco

Abstract: Many-core tiled CMP proposals often assume a partially shared last level cache (LLC) since this provides a good compromise between access latency and cache utilization. In this paper, we propose a novel way to map memory addresses to LLC banks that takes into account the average distance between the banks and the tiles that access them. Contrary to traditional approaches, our mapping does not group the tiles in clusters within which all the cores access the same bank for the same addresses. Instead, two neighb… Show more

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1

Citation Types

0
0
0

Year Published

2012
2012
2014
2014

Publication Types

Select...
3
2

Relationship

1
4

Authors

Journals

citations
Cited by 5 publications
(1 citation statement)
references
References 21 publications
0
0
0
Order By: Relevance
“…NoC-aware cache design [43] introduces a first-touch data placement policy, a migration policy that moves each block to its most frequent sharer and a replacement policy that is biased towards retaining shared blocks and replacing private ones. DAPSCO [78] consists on a distance-aware cache organization that minimizes the average distance travelled by cache requests. In our system the L2 cache is centralized instead of distributed, since the number of cores in a mobile GPU is much smaller than what is assumed in a many-core system due to power constraints.…”
Section: Chapter 1 Introductionmentioning
confidence: 99%
“…NoC-aware cache design [43] introduces a first-touch data placement policy, a migration policy that moves each block to its most frequent sharer and a replacement policy that is biased towards retaining shared blocks and replacing private ones. DAPSCO [78] consists on a distance-aware cache organization that minimizes the average distance travelled by cache requests. In our system the L2 cache is centralized instead of distributed, since the number of cores in a mobile GPU is much smaller than what is assumed in a many-core system due to power constraints.…”
Section: Chapter 1 Introductionmentioning
confidence: 99%