jpff wrote:
> There is an "improved" sore sorter in the file Engine/smoothsort.c
> which is a straight replacement for Engine/sort.c

Incredible improvement!!

I made a simple score with random p-fields composed of 900,000 events.
The new sort finished the job in about 10 seconds.  I tried my old binary,
and waited, and waited, browsed the web, checked my email.  I gave up.
I took the size down an order of magnitude to 90,000 events.  The old
binary took almost two minutes to sort that, while the new sorter took
about exactly one second!

Thanks John, you have made my life quite a bit nicer.

Now for the bug.  One line in my 90,000 line score was left
unsorted.  I found it at the very end of the file.  It must
be one of those thing where the pointer arithmetic is not
quite right.  I've bundled the source score, and the resulting
'slow' and 'fast' .srt files (Ha! a use for --keep-sorted-score) here:

http://tobiah.org/scores.zip

By the way, I checked my 900,000 line version of the
score, and the same thing was true: the very last line
was out of sequence.

Once this little bug is fixed, I'd say that you should make
the new sorter the default in the build.

Thanks,

Tobiah


Send bugs reports to the Sourceforge bug tracker
            https://sourceforge.net/tracker/?group_id=81968&atid=564599
Discussions of bugs and features can be posted here
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"

Whoops!  This is almost certainly because there is not an e at the end of
the score which I stupidly assumed was there.  Willk change.  Glad it
(otherwise) worked!
==John ff

> jpff wrote:
>> There is an "improved" sore sorter in the file Engine/smoothsort.c
>> which is a straight replacement for Engine/sort.c
>
> Incredible improvement!!
>
> I made a simple score with random p-fields composed of 900,000 events.
> The new sort finished the job in about 10 seconds.  I tried my old binary,
> and waited, and waited, browsed the web, checked my email.  I gave up.
> I took the size down an order of magnitude to 90,000 events.  The old
> binary took almost two minutes to sort that, while the new sorter took
> about exactly one second!
>
> Thanks John, you have made my life quite a bit nicer.
>
> Now for the bug.  One line in my 90,000 line score was left
> unsorted.  I found it at the very end of the file.  It must
> be one of those thing where the pointer arithmetic is not
> quite right.  I've bundled the source score, and the resulting
> 'slow' and 'fast' .srt files (Ha! a use for --keep-sorted-score) here:
>
> http://tobiah.org/scores.zip
>
> By the way, I checked my 900,000 line version of the
> score, and the same thing was true: the very last line
> was out of sequence.
>
> Once this little bug is fixed, I'd say that you should make
> the new sorter the default in the build.
>
> Thanks,
>
> Tobiah
>
>
> Send bugs reports to the Sourceforge bug tracker
>             https://sourceforge.net/tracker/?group_id=81968&atid=564599
> Discussions of bugs and features can be posted here
> To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe
> csound"
>
>
>
>




Send bugs reports to the Sourceforge bug tracker
            https://sourceforge.net/tracker/?group_id=81968&atid=564599
Discussions of bugs and features can be posted here
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"

> Now for the bug.  One line in my 90,000 line score was left
> unsorted.  I found it at the very end of the file.

and now fixed in CVS
==John ff



Send bugs reports to the Sourceforge bug tracker
            https://sourceforge.net/tracker/?group_id=81968&atid=564599
Discussions of bugs and features can be posted here
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"

>> Now for the bug.  One line in my 90,000 line score was left
>> unsorted.  I found it at the very end of the file.
>
> and now fixed in CVS
> ==John ff

It's perfect now, Thanks!

Tobiah


Send bugs reports to the Sourceforge bug tracker
            https://sourceforge.net/tracker/?group_id=81968&atid=564599
Discussions of bugs and features can be posted here
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"

I was just reading a series of posts beginning in January and ending  
with the note below in March. In that relatively short period of time  
it seems that csound was improved immeasurably.

This is frankly awesome. I'm so glad to be part of this community, if  
only generally an onlooker.

best,
Lou Cohen


On Mar 13, 2010, at 12:05 PM, Tobiah wrote:

>>> Now for the bug.  One line in my 90,000 line score was left
>>> unsorted.  I found it at the very end of the file.
>>
>> and now fixed in CVS
>> ==John ff
>
> It's perfect now, Thanks!
>
> Tobiah
>
>
> Send bugs reports to the Sourceforge bug tracker
>            https://sourceforge.net/tracker/?group_id=81968&atid=564599
> Discussions of bugs and features can be posted here
> To unsubscribe, send email sympa@lists.bath.ac.uk with body  
> "unsubscribe csound"
>



Send bugs reports to the Sourceforge bug tracker
            https://sourceforge.net/tracker/?group_id=81968&atid=564599
Discussions of bugs and features can be posted here
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound"