Re: what's it worth to write a short program for polynomial multiplication?
- From: rjf <fateman@xxxxxxxxx>
- Date: Mon, 2 Jun 2008 21:18:09 -0700 (PDT)
On Jun 2, 5:41 pm, TimDaly <d...@xxxxxxxxxxxxxxxxxxx> wrote:
......
There are 58 * operators in Axiom, many of which can by automatically
specialized for the polynomial type in question. Thus,
x * y
is the shortest, valid way to efficiently multiply two polynomials in
the Spad (scratchpad, Axiom's native language).
Well, there may still be decisions to be made that have to do with the
actual data, as opposed to the declared category.
For example, one could have data that is declared to be a sparse
polynomial and is stored that way, but actually has entirely filled
in. If the user knew that, a coercion might be introduced. But if the
user did not, and the program just keyed off the categories of the
inputs, the algorithm used would not be especially efficient. Should
the "*" operator do the work necessary to test the data at runtime (as
opposed to the "type" of the data)?
It would be one way of trying to use the "most efficient" algorithm at
the cost of slowing everything down just a little :)
Tim Daly
Axiom Lead Developer
.
- Follow-Ups:
- References:
- what's it worth to write a short program for polynomial multiplication?
- From: rjf
- Re: what's it worth to write a short program for polynomial multiplication?
- From: Herman Rubin
- Re: what's it worth to write a short program for polynomial multiplication?
- From: rjf
- Re: what's it worth to write a short program for polynomial multiplication?
- From: TimDaly
- what's it worth to write a short program for polynomial multiplication?
- Prev by Date: 'Manipulate' in Documentation Centre slow to load (sometimes fails)
- Next by Date: Re: Manipulate - discrete values with names + images.
- Previous by thread: Re: what's it worth to write a short program for polynomial multiplication?
- Next by thread: Re: what's it worth to write a short program for polynomial multiplication?
- Index(es):
Relevant Pages
|
