Cache-Oblivious Algorithms (Extended Abstract)

Cache-Oblivious Algorithms (Extended Abstract) (1999)

Cached

Download Links

by Matteo Frigo , Charles E. Leiserson , Harald Prokop , Sridhar Ramachandran , Z W(l

Venue:	In Proc. 40th Annual Symposium on Foundations of Computer Science
Citations:	10 - 1 self

Versions

Version History

Metadata Version 3

User correction supplied by SystemCorrections

Datum	Value	Source
TITLE	Cache-Oblivious Algorithms (Extended Abstract)	user correction - Legacy Corrections
AUTHOR NAME	Matteo Frigo	SVM HeaderParse 0.1
AUTHOR AFFIL	MIT Laboratory for Computer Science	SVM HeaderParse 0.2
AUTHOR ADDR	545 Technology Square, Cambridge, MA 02139	SVM HeaderParse 0.1
AUTHOR NAME	Charles E. Leiserson	SVM HeaderParse 0.1
AUTHOR AFFIL	MIT Laboratory for Computer Science	SVM HeaderParse 0.2
AUTHOR ADDR	545 Technology Square, Cambridge, MA 02139	SVM HeaderParse 0.1
AUTHOR NAME	Harald Prokop	SVM HeaderParse 0.1
AUTHOR AFFIL	MIT Laboratory for Computer Science	SVM HeaderParse 0.2
AUTHOR ADDR	545 Technology Square, Cambridge, MA 02139	SVM HeaderParse 0.1
AUTHOR NAME	Sridhar Ramachandran	SVM HeaderParse 0.1
AUTHOR AFFIL	MIT Laboratory for Computer Science	SVM HeaderParse 0.2
AUTHOR ADDR	545 Technology Square, Cambridge, MA 02139	SVM HeaderParse 0.1
AUTHOR NAME	Z W(l	SVM HeaderParse 0.1
AUTHOR AFFIL	2	SVM HeaderParse 0.2
ABSTRACT	This paper presents asymptotically optimal algorithms for rectangular matrix transpose, FFT, and sorting on computers with multiple levels of caching. Unlike previous optimal algorithms, these algorithms are cache oblivious: no variables dependent on hardware parameters, such as cache size and cache-line length, need to be tuned to achieve optimality. Nevertheless, these algorithms use an optimal amount of work and move data optimally among multiple levels of cache. For a cache with size Z and cache-line length L where Z = W(L 2 ) the number of cache misses for an m \Theta n matrix transpose is Q(1 +mn=L). The number of cache misses for either an n-point FFT or the sorting of n numbers is Q(1+ (n=L)(1 + log Z n)). We also give an Q(mnp)-work algorithm to multiply an m \Theta n...	user correction - Legacy Corrections
YEAR	1999	INFERENCE
VENUE	In Proc. 40th Annual Symposium on Foundations of Computer Science	INFERENCE
VENUE TYPE	CONFERENCE	INFERENCE
PAGES	285--397	INFERENCE
CITATIONS	37 found	ParsCit 1.0