| On Sat, Jan 23, 2010 at 7:20 AM, wrote:
> Rather than reengineering the sorting, wouldn't it be a lot easier to
> implement the original poster's suggestion to allow the specification
> of alternate input/output names for the sorted score file?
I didn't realize that a vastly superior sorting algorithm was available.
I vote that this sorting inefficiency be corrected as soon as possible.
In my original post, I only asked for a flag to rename the score, because
I was largely ignorant as to the csound sort method that was being used.
I have a recent processor, and yet it took csound over 80 minutes to
sort a score that had less than a half million lines in it!
For comparison, I tried writing a python program that generates
10 million random numbers, and sorts them. Never mind the generation time,
the entire program completed in just over 20 seconds!:
import random
a = []
print "Entering Ten Million Loop"
for x in range(10000000L):
a.append(random.random())
print "Done with array creation. Starting Sort."
a.sort()
print "Done with sorting"
print len(a)
So gentlemen, please, let's regard this as a very serious issue;
csound ought to be able to handle scores of very large lengths.
I need it to be able to handle such scores. Concept: using a score generation
program for grain-type synthesis.
Thanks as always for all that contribute to this awesome piece of software.
Tobiah
Send bugs reports to this list.
To unsubscribe, send email sympa@lists.bath.ac.uk with body "unsubscribe csound" |