org.ejml.alg.dense.misc
Class TransposeAlgs
- java.lang.Object
-
- org.ejml.alg.dense.misc.TransposeAlgs
-
public class TransposeAlgs extends java.lang.ObjectLow level transpose algorithms. No sanity checks are performed. Take a look at BenchmarkTranspose to see which one is faster on your computer.
-
-
Constructor Summary
Constructors Constructor and Description TransposeAlgs()
-
Method Summary
All Methods Static Methods Concrete Methods Modifier and Type Method and Description static voidblock(RowD1Matrix64F A, RowD1Matrix64F A_tran, int blockLength)Performs a transpose across block sub-matrices.static voidsquare(RowD1Matrix64F mat)In-place transpose for a square matrix.static voidstandard(RowD1Matrix64F A, RowD1Matrix64F A_tran)A straight forward transpose.
-
-
-
Method Detail
-
square
public static void square(RowD1Matrix64F mat)
In-place transpose for a square matrix. On most architectures it is faster than the standard transpose algorithm, but on most modern computers it's slower than block transpose.- Parameters:
mat- The matrix that is transposed in-place. Modified.
-
block
public static void block(RowD1Matrix64F A, RowD1Matrix64F A_tran, int blockLength)
Performs a transpose across block sub-matrices. Reduces the number of cache misses on larger matrices. *NOTE* If this is beneficial is highly dependent on the computer it is run on. e.g: - Q6600 Almost twice as fast as standard. - Pentium-M Same speed and some times a bit slower than standard.- Parameters:
A- Original matrix. Not modified.A_tran- Transposed matrix. Modified.blockLength- Length of a block.
-
standard
public static void standard(RowD1Matrix64F A, RowD1Matrix64F A_tran)
A straight forward transpose. Good for small non-square matrices.- Parameters:
A- Original matrix. Not modified.A_tran- Transposed matrix. Modified.
-
-
DMelt 3.0 © DataMelt by jWork.ORG