la.env(NB=.laenv$NB, NBMIN=.laenv$NBMIN, NX=.laenv$NX, NS=.laenv$NS, NXSVD=.laenv$NXSVD, MAXB=.laenv$MAXB)
NBMIN
,
an unblocked method is used.
The value of
NBMIN
is irrelevant
if
NB=1
.
NX
, an unblocked method is used.
The value of
NX
is irrelevant
if
NB=1
.
la.env
,
with missing parameters replaced by their default values.
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.
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.
Anderson, E., et al. (1994), LAPACK Users' Guide, 2nd edition. SIAM, Philadelphia.
la.env(NB=4, NBMIN=2, NX=2)