Re: Want to use results by Parallel Computing
- To: mathgroup at smc.vnet.net
- Subject: [mg26996] Re: Want to use results by Parallel Computing
- From: Jens-Peer Kuska <kuska at informatik.uni-leipzig.de>
- Date: Tue, 30 Jan 2001 23:22:17 -0500 (EST)
- Organization: Universitaet Leipzig
- References: <9563bt$9f@smc.vnet.net>
- Sender: owner-wri-mathgroup at wolfram.com
Hi,
ParallelEvaluate[Do[c[i] = i * (i+2), {i, 1000}]]
set the definitions for c[] up on several machines.
and
ParallelEvaluate[Do[c[i], {i, 1000}]]
will pass the values of the function definitions back
to the master machine.
Regards
Jens
>
>
"Toshiyuki (Toshi) Meshii" wrote:
>
> Hello,
>
> Some strange things about parallel computing.
>
> When I execute the following on a single machine, there is no problem.
>
> Do[c[i] = i * (i+2), {i, 1000}]
>
> I can surely obtain c[5] or any of the elements.
>
> c[5]
> 35
>
> However, when I do the same thing by ParallelEvaluate, executing time
> becomes shorter, but I cannot access to the results.
>
> ParallelEvaluate[Do[c[i] = i * (i+2), {i, 1000}]]
>
> That is, I cannot get the concrete value of c[5] on the master machine.
>
> c[5]
> c[5]
>
> Mathematica just returns c[5] symbolically.
>
> What is happening?
> I hope some one can give me advice on obtaining the results, such as
> c[5]=35.
>
> -Toshi