On Fri, Feb 03, 2012 at 03:53:10AM -0800, Martin Baker wrote:
> relative performance issues so I will leave you to choose whether to
> substitute this:
well I did some measurements which indicate that you code is faster,
the difference increasing with the size of the matrices.
Profiling indicates that horizConcat and vertConcat seem to add
avoidable overhead. Don't know whether it can be improved,
that's perhaps the price to pay for compactness.
Trying map from MATCAT2,
with the function aij +-> aij*B fails with the error
"Function: 0 : () -> % is missing from domain: Matrix(Float)"
here is the test (debian amd64):
g:=normal(0,1)$RFDIST
randmat(n) ==
matrix [[g() for i in 1..n] for j in 1..n]
for k in 1..100 repeat A:=tensorProduct(randmat(10),randmat(10));
Type: Void
Time: 3.36 (EV) = 3.36 sec
for k in 1..100 repeat A:=kroneckerProduct(randmat(10),randmat(10));
Type: Void
Time: 0.004 (IN) + 3.18 (EV) + 0.01 (OT) = 3.20 sec
for k in 1..100 repeat A:=tensorProduct(randmat(20),randmat(10));
Type: Void
Time: 14.71 (EV) = 14.71 sec
for k in 1..100 repeat A:=kroneckerProduct(randmat(20),randmat(10));
Type: Void
Time: 11.67 (EV) = 11.67 sec
for k in 1..100 repeat A:=tensorProduct(randmat(30),randmat(30));
Type: Void
Time: 386.77 (EV) = 386.77 sec
for k in 1..100 repeat A:=kroneckerProduct(randmat(30),randmat(30));
Type: Void
Time: 248.20 (EV) = 248.20 sec
best,
Franz
--
You received this message because you are subscribed to the Google Groups
"FriCAS - computer algebra system" group.
To post to this group, send email to [email protected].
To unsubscribe from this group, send email to
[email protected].
For more options, visit this group at
http://groups.google.com/group/fricas-devel?hl=en.
