Re: CALCULATION TIME
- To: mathgroup at smc.vnet.net
- Subject: [mg21427] Re: CALCULATION TIME
- From: Jens-Peer Kuska <kuska at informatik.uni-leipzig.de>
- Date: Fri, 7 Jan 2000 00:20:38 -0500 (EST)
- Organization: Universitaet Leipzig
- References: <851ek8$4b5@smc.vnet.net>
- Sender: owner-wri-mathgroup at wolfram.com
Hi,
whats wrong ? The Do[]-Loops ? You should look into your expression
to express the operations with highlevel commands like Dot[],
RotateLeft[],
ListConvolve[], Fold[] ...
I use rather seldom Do[]'s and For[]'s and I have never missed it.
Functional
operators like Map[], Apply[] are typical faster.
You supply not the information to give you more hints.
Regards
Jens
Blimbaum Jerry DLPC wrote:
>
> I have some code performing (roughly) the following Do Loops:
>
> Do[.........
> Do[x[[m,n+1]] = x[[m,n]] + B* y[[n+m-1]]*e[[n]]
>
> ----------------------------------------------,
>
> eps+Sum[y[[n-q]]^2,{q,0,M-1}]
> {m,M} ],
>
> {n,M,Length[y]}]
>
> where y, e are lists (y is data that contains about 30000 elements),
> etc. Running a data set takes about 10.5 secs. I wanted to shorten the
> time but couldnt figure out how to write it either in Nest or Compile form
> so I went thru it piece meal to see if I could find a culprit and from this
> I noticed if I replace the Sum term in the denominator by a number, lets say
> 1, and then perform the calculations it only takes about 2.5 seconds! I
> did a sample Sum[expr^2] with a similar Do loop separately and it takes no
> time at all.
>
> What is wrong?
>
> thanks. Jerry Blimbaum NSWC Panama City, Fl