Patent · US Active

System and method for memory bandwidth friendly sorting on multi-core architectures

US8463820B2 · kind B2 · utility

6Cited by
5References
28Claims
0Family size

Assignee

Inventors

Key dates

Filing dateMay 26, 2009
Grant dateJun 11, 2013
Priority date
Expiry dateDec 4, 2030

Classification

  • Technology area (CPC G)Physics
  • CPC primaryG06F12/0802
  • WIPO fieldComputer technology
  • WIPO sectorElectrical engineering

Abstract

In some embodiments, the invention involves utilizing a tree merge sort in a platform to minimize cache reads/writes when sorting large amounts of data. An embodiment uses blocks of pre-sorted data residing in “leaf nodes” residing in memory storage. A pre-sorted block of data from each leaf node is read from memory and stored in faster cache memory. A tree merge sort is performed on the nodes that are cache resident until a block of data migrates to a root node. Sorted blocks reaching the root node are written to memory storage in an output list until all pre-sorted data blocks have been moved to cache and merged upward to the root. The completed output list in memory storage is a list of the fully sorted data. Other embodiments are described and claimed.

Source: USPTO / EPO open patent data. Objective bibliographic and citation counts.