2005 International Conference on Computer Design
DOI: 10.1109/iccd.2005.65
|View full text |Cite
|
Sign up to set email alerts
|

Implementing caches in a 3D technology for high performance processors

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
2
1
1
1

Citation Types

0
50
0

Publication Types

Select...
5
2
2

Relationship

0
9

Authors

Journals

citations
Cited by 92 publications
(50 citation statements)
references
References 19 publications
0
50
0
Order By: Relevance
“…Loh proposes optimizations to 3D DRAM that result in 1.75x speedup over prior 3D-DRAM approaches, and Loh also proposes a L2 miss handling architecture that achieves an extra 17.8% performance improvement [14]. Puttaswamy and Loh show that a 3D-partitioned cache can reduce latency by 21.5%, reduce energy consumption by 30.9%, and increase IPC by 12% [22].…”
Section: Motivation For 3d Securitymentioning
confidence: 99%
“…Loh proposes optimizations to 3D DRAM that result in 1.75x speedup over prior 3D-DRAM approaches, and Loh also proposes a L2 miss handling architecture that achieves an extra 17.8% performance improvement [14]. Puttaswamy and Loh show that a 3D-partitioned cache can reduce latency by 21.5%, reduce energy consumption by 30.9%, and increase IPC by 12% [22].…”
Section: Motivation For 3d Securitymentioning
confidence: 99%
“…Many have focused on improving single-core performance and power [2,7,26,44]. Some of this attention has focused on implementing a cache in 3D [25,29,40], even in the context of a multi-core NUCA layout [16]. Few studies have considered using additional dies entirely for SRAM or DRAM [2,17,18].…”
Section: Related Workmentioning
confidence: 99%
“…In MLBS, the process of building active device layers is repeated on a single wafer to compose multiple dies before processing all the metal routing layers. 3D integration techniques enable mixing dissimilar process technologies, such as highspeed CMOS and high-density DRAM by using MLBS, which make it possible to stack many heterogeneous dies [25,32]. While MLBS require changes in the manufacturing process, die-bonding 3D integration techniques insert metal vias to bond the two planar dies based on the conventional 2D manufacturing process.…”
Section: Related Workmentioning
confidence: 99%
“…For the 3D processor, we partition and stack one core and half of the L2 cache on each of the die [32,42]. Though there are alternatives to stack the cores and L2 caches, we consider the 3D processor which has one core and half of the L2 cache on each of the die.…”
Section: D Processor Incorporated With Thementioning
confidence: 99%