Re: Re: Split in Mathematica 2.2
- To: mathgroup at smc.vnet.net
- Subject: [mg12980] Re: [mg12822] Re: [mg12777] Split in Mathematica 2.2
- From: Andrzej Kozlowski <andrzej at tuins.ac.jp>
- Date: Sun, 28 Jun 1998 02:52:04 -0400
- References: <6msqgp$i7j$3@dragonfly.wolfram.com>
- Sender: owner-wri-mathgroup at wolfram.com
For those interested in this topic here is a summary of our conclusions. The really intriguing thing about Split in Mathematica 3.0 is its apparent non-linear scaling on the Macintosh. As the result, the fastest of our user defined versions, which scales linearly, beats the built in function for speed for large lists. (The exact point where the user defined verion overtakes the built-in one varies according to processor, but occurs earliest on G3 Macs. If any one would like to test this behaviour I can provide a notebook fro this purpose). We have contacted technical support. They confirmed that the algorithm contains no non-linearities, and is the same on all platforms. On other platforms Split seems to scale linearly and easily beats all user defined versions. We were told that Mathematica development team is investigating this matter. My personal opinion is that this has something to do with the compiler. If so it would be interesting to test scaling of other built in functions. Andrzej Kozlowski