Re: Is there something quicker than textscan?

[PhilipNienhuis]

Jordi Gutiérrez Hermoso-2 wrote
> 
> On 19 March 2012 07:44, PhilipNienhuis <pr.nienhuis@> wrote:
>> I have a patch in mind (and tried in a debug situation) that will
>> significantly speed up your problem (though not as fast as ML), but
>> that will have to wait until (1) I got time to implement it and (2)
>> release of Octave 3.8. Or maybe 3.6.2 if it the core developers
>> agree that the poor performance is a bug (I'd say so but who am I).
> 
> The stable release is supposed to be conservative. It has to be, if it
> is to be stable. Patching a non-critical performance bug is not
> conservative, because that may introduce unforeseen bugs of its own
> and not achieve the desired stability. Is the performance bad enough
> to make the function unusable in all or most situations? In that case,
> it's a patch for the stable branch. If not, it can wait until the
> northern hemisphere summer, which is the rough expected release date
> of the 3.8 version.
> 

Patch is ready - see:

https://savannah.gnu.org/patch/?7748

but anyway you're quite right, Jordi. 
I thought it was easy (I had a big speed up working very soon) but then I
hit several unforeseen issues (a situation not unlike when I started fixing
the strread/textread/textscan trio a year ago. Took me quite a while).
Hopefully I haven't introduced new bugs (though I have tested it quite
intensively and squashed a few yet unnoticed bugs on the way).

Using an example resembling the one from the OP I get a 70+ times speed
improvement. Yet a binary solution (dlmread) is again 10 X faster.

Philip


--
View this message in context: 
http://octave.1599824.n4.nabble.com/Is-there-something-quicker-than-textscan-tp4483991p4490731.html
Sent from the Octave - General mailing list archive at Nabble.com.