Re: Programming skills required by mathematicians?

In article <bc59a$4863929e$82a1e228$2537@[EMAIL PROTECTED]
>,
Han de Bruijn  <Han.deBruijn@[EMAIL PROTECTED]
> wrote:
>toni.lassila@[EMAIL PROTECTED]
 wrote:

>> On Jun 26, 9:34 am, Han de Bruijn <Han.deBru...@[EMAIL PROTECTED]
> wrote:

>>>Herman Rubin wrote:

>>>>Computers are essentially a fixed program, the
>>>>hardware, which then has a small initialized amount
>>>>of software, quite variable, which enables it to
>>>>try to do what the user wants.  Keep this in mind;
>>>>understand what the hardware can do, which is
>>>>different on different machines, and think in those
>>>>terms, not the current poorly designed languages.

>>>Guess you're still thinking of an IBM mainframe with
>>>APL on it ?

>> No, with parallel computing it is now even more im****tant
>> to take into account the hardware implementation.

>Ah, parallel computing ! I guess Herman Rubin didn't talk about that.

I included this.  As I recall, I did say that timing
considerations might not be covered by the simulation.

>> A relatively naive but effectively vectorized algorithm
>> can beat that highly tuned F77 code from decades ago that
>> loops all over the place.

>Vectorized ? I thought you were talking about parallel .. ?

I am quite aware of the problems on vector and parallel
machines.  I know how to modify many rather efficient
procedures for generating non-uniform random numbers to
take this into account, and these modifications, while
somewhat complicated, seem to be candidates for machines
which are neither vectorized or parallel.



-- 
This address is for information only.  I do not claim that these views
are those of the Statistics Department or of Purdue University.
Herman Rubin, Department of Statistics, Purdue University
hrubin@[EMAIL PROTECTED]
         Phone: (765)494-6054   FAX: (765)494-0558