Set Tuning Parameters for Linear Algebra Computations

DESCRIPTION:

Sets values of tuning parameters within LAPACK.

USAGE:

la.env(NB=.laenv$NB, NBMIN=.laenv$NBMIN, NX=.laenv$NX, NS=.laenv$NS,
       NXSVD=.laenv$NXSVD, MAXB=.laenv$MAXB)

REQUIRED ARGUMENTS:

NB
optimal blocksize.
NBMIN
minimum block size. If the usable block size is less than NBMIN, an unblocked method is used. The value of NBMIN is irrelevant if NB=1 .
NX
crossover point. In a block method, when blocks reach dimension less than NX, an unblocked method is used. The value of NX is irrelevant if NB=1.
NS
number of shifts for nonsymmetric eigenvalue computations.
NXSVD
crossover point for the singular-value computation. When reducing an m by n matrix to bidiagonal form, if max(m,n)/min(m,n) exceeds this value, the matrix is first reduced to triangular form by QR factorization.
MAXB
crossover point within nonsymmetric eigenvalue computations.

VALUE:

returns the list of parameters used in the call to la.env , with missing parameters replaced by their default values.

SIDE EFFECTS:

The corresponding parameters within Lapack are set to the values supplied in la.env. The following is from Anderson et al. (1994): The range of problem sizes needed to determine the optimal block size or crossover point is machine dependent. For algorithms that require a crossover point, it is best to start by finding the best block size with the crossover point set to 0, and then locate the point at which blocked algorithm begins to outperform then unblocked algorithm using this block size.r The best crossover point will be somewhat smaller than the latter value. By experimenting with small values, it should be straightforward to choose NBMIN, the smallest block size that gives improvement over an unblocked method.

NOTE:

Currently only a few of the linear algebra functions interact with these tuning parameters. Functions that do use these value have a referenced to la.env in their help file.

REFERENCES:

Anderson, E., et al. (1994), LAPACK Users' Guide, 2nd edition. SIAM, Philadelphia.

SEE ALSO:

EXAMPLES:

la.env(NB=4, NBMIN=2, NX=2)