Re: [Changeset] Re: Faster Array transpose

[dbateman]

John W. Eaton wrote:
> 
> That's a lot of work for a fairly modest improvement...
> 

Well transpose and hermitian operators are used rather often and so even
modest gains are useful. Also the gain is about a factor of 2 for matrix
sizes divisible by 8. In any case the reason I started this was to move more
methods into the Array class itself such that the float array classes could
be simplified, so the speed gain in that sense in secondary. Still need a
little work there before I'd suggest you take my code, but its almost there.



> Is there a ChangeLog entry for this?  With that, I'd apply the change
> (but skipping the diffs for the float matrix files I don't have yet).
> 

Humm, I was sure I wrote one, but it seems not. I;ll send another changeset
with one later.

All the float stuff is on my respository at hg.dbateman.org, starting at
changeset 8002.



> Also, how about some tests for this since the algorithm is now more
> complex?
> 

Sure ok, but there are lots of places where make check already uses the
transpose and that was useful in debugging the code.. In any case I'll
supply some specific tests later as well.

D.

-- 
View this message in context: 
http://www.nabble.com/Faster-Array-transpose-tp16992703p17090631.html
Sent from the Octave - Maintainers mailing list archive at Nabble.com.