Proceedings of the Twenty-Second International Conference on Architectural Support for Programming Languages and Operating Syst 2017
DOI: 10.1145/3037697.3037749
|View full text |Cite
|
Sign up to set email alerts
|

Moonwalk

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1

Citation Types

0
2
0

Year Published

2018
2018
2021
2021

Publication Types

Select...
6
1

Relationship

1
6

Authors

Journals

citations
Cited by 23 publications
(2 citation statements)
references
References 48 publications
0
2
0
Order By: Relevance
“…LPDDR at 400 mbps per pin is about 5× off in bandwidth from more modern DDR; but the latency is close, maybe 1.3× off; so for processor cores (like the RISC-V processors described earlier), which are latency-bound, it is reasonable, for accelerators, which are often bandwidth-bound it is less so. Typical server accelerator chips will need PCI-E, which [19].…”
Section: Open Source Ipmentioning
confidence: 99%
“…LPDDR at 400 mbps per pin is about 5× off in bandwidth from more modern DDR; but the latency is close, maybe 1.3× off; so for processor cores (like the RISC-V processors described earlier), which are latency-bound, it is reasonable, for accelerators, which are often bandwidth-bound it is less so. Typical server accelerator chips will need PCI-E, which [19].…”
Section: Open Source Ipmentioning
confidence: 99%
“…f This true for the 16/14nm node size. Lithography cost are by far the biggest cost component of the manufacturing NRE 21. Other costs include labor and design tools, as well as IP licensing.…”
mentioning
confidence: 99%