clusteringfittingintegrationinterpolationjavaleast-squaresmathoptimizationpolynomialsolverstattransforms
e01294b158
described in Siddhartha Chatterjee, Alvin R. Lebeck, Praveen K. Patnala and Mithuna Thottehodi paper "Recursive Array Layout and Fast Matrix Multiplication". As of 2009-02-13, this implementation does not work! The padding at left and bottom sides of the matrix should be cleared after some operations like scalerAdd and is not. Also there is a limitation in the multiplication that can only process matrices with sizes similar enough to have the same power of two number of tiles in all three matrices A, B and C such that C = A*B. These parts have not been fixed since the performance gain with respect to DenseRealMatrix are not very important, and the numerical stability is poor. This may well be due to a bad implementation. This code has been put in the experimental directory for the record, putting it into production would require solving all these issues. git-svn-id: https://svn.apache.org/repos/asf/commons/proper/math/trunk@744126 13f79535-47bb-0310-9956-ffa450edef68 |
||
---|---|---|
src | ||
LICENSE.txt | ||
NOTICE.txt | ||
PROPOSAL.html | ||
RELEASE-NOTES.txt | ||
build.properties.sample | ||
build.xml | ||
checkstyle.xml | ||
doap_math.rdf | ||
findbugs-exclude-filter.xml | ||
license-header.txt | ||
maven.xml | ||
pom.xml | ||
project.properties | ||
project.xml | ||
release-notes.jsl | ||
test-jar.xml | ||
testOnly.sh |