TechWeb Digital Library

High-Performance Linear Algebra Algorithms Using New Generalized Data Structures for Matrices

Source: IBM
Date: January 2008
Type: White Paper
Rating: (0)

Overview: This research paper describes several novel data formats for dense linear algebra and some simple, novel algorithms that utilize these new data structures. The paper relies on a heuristic that is the key factor governing performance on processors with deep memory hierarchies, namely blocking or tiling. To be able to use this heuristic, the researchers have made use of simple fact from linear algebra.


Click here to download now

View all content from this source

Not what you’re looking for? Search again
Go Advanced »
Email Alert

Receive an email alert whenever new content is added to the Business Intelligence section of the TechWeb Digital Library

More Business Intelligence Resources

Why BI is Ripe - NOW - For Businesses of Any Size
Oracle's range of offerings to mid-size and emerging companies reflects its vision that BI and EPM solutions can...

Impact Business Performance with BPM
To maximize your chances for success, you need the ability to continually optimize, streamline and align business processes to meet changing business needs for greater performance. Business process management (BPM) from IBM can help.

Upcoming Webcasts

More On Business Intelligence