1994
DOI: 10.1109/92.311644
|View full text |Cite
|
Sign up to set email alerts
|

Determining objective functions in systolic array designs

Help me understand this report

Search citation statements

Order By: Relevance

Paper Sections

Select...
1
1
1
1

Citation Types

0
3
0
1

Year Published

1997
1997
2011
2011

Publication Types

Select...
6
2

Relationship

0
8

Authors

Journals

citations
Cited by 12 publications
(4 citation statements)
references
References 10 publications
0
3
0
1
Order By: Relevance
“…Among them only one, belonging to Class I has minimal number of PEs for the given matrix dimensions. This can be proved according to the following corollary as given in [23]. …”
Section: Modification Of the Design Proceduresmentioning
confidence: 96%
“…Among them only one, belonging to Class I has minimal number of PEs for the given matrix dimensions. This can be proved according to the following corollary as given in [23]. …”
Section: Modification Of the Design Proceduresmentioning
confidence: 96%
“…One from hexagonal group SAs, which also belong to planar SAs and that have all components of projection direction which have only value ±1, only the SA given for projection direction µ = [1 1 1] T enables the calculation of high dependability, therefore it has t p = 3. This SA has been much studied in the literature as one fault tolerant design (see [5,7,10]). …”
Section: Case (I)mentioning
confidence: 99%
“…The communication links between the PEs are obtained by mapping dependency matrix Q, defined in (3), using the same matrix S. However, the obtained BLSA is not optimal with respect to the number of PEs for a given problem size. According to Theorem defined in [13], it can be concluded that number of PEs in this BLSA is Ω = 2n − 1. In order to obtain SA with optimal number of PEs, i.e.…”
Section: The Systolic Algorithmmentioning
confidence: 99%