RE: Mathematica vs. Executable code
- To: mathgroup at smc.vnet.net
- Subject: [mg15973] RE: Mathematica vs. Executable code
- From: "Ersek, Ted R" <ErsekTR at navair.navy.mil>
- Date: Fri, 19 Feb 1999 03:26:57 -0500
- Sender: owner-wri-mathgroup at wolfram.com
Charles replied to my message with:
I've been looking for some gossip about new version of mma (4.0?). It looks
like you heard something ('packed arrays'). Anything else? And when are they
going to release that?
_____________________
I attended Mathematica Conference 98 last June where they had computers with
Version 4.0 for us to use. This version includes 'Packed Arrays' which are
efficient ways of handling vectors, matrices, (tensors too I think) when
every element is a machine number. For the most part the user doesn't have
to think about whether something is or isn't a Packed Array. The kernel
takes care of it automatically, and does things faster. As I recall it
often runs 5 times faster than Version 3.0 when you do things with a large
Packed Array.
Version 4.0 will also include:
- Spelling correction (like in a word processor)
- Nice automatic indenting which will make it easy to keep track
of nested brackets [], braces {}, and parentheses () when programming.
- New features to help you keep track of unmatched brackets [],
braces {}, and parentheses ().
- Some new math features, but I forget what they are.
I can only guess about when it will be released. I got reports that it went
into Beta test a few months ago.
________________________________________
I started thinking about comparing the performance of Version 4.0 with
compiled programs. It seems one would only need to write a C program to
multiply a long list of real numbers. You could then see how long it takes
Mathematica verses the C program to multiply the numbers. Then you could do
the same for adding all the numbers, and computing say Log[Abs[_]] on each
number. Even I can write such a simple C program. It should be
interesting.
Regards,
Ted Ersek