Re: [boost] Re: matrix library ?

From: Jeremy Siek (jsiek_at_[hidden])
Date: 2001-03-20 10:40:07

Hmm, I think what Oleg calls "lazy matrices" is something that I would
normally just bundle into what I think of as expression templates...
though there's no point in arguing over the terminology. Either way, its
an important concept. Creating various transformed "views" of a matrix and
choosing the right time to evaluate is certainly necessary.

Here's a paper I wrote about the MTL v3.0 prototype concerning lazy/eager
evaluation in the context of matrix expressions:

http://www.lsc.nd.edu/~jsiek/mxtl.html

Cheers,
Jeremy

On Tue, 20 Mar 2001, Joel Young wrote:

jdy> 
jdy> Jeremy Siek wrote:
jdy> > These lazy matrices sound and aweful lot like expression
jdy> > templates. Is there any real difference?
jdy> 
jdy> I think so.  I just looked at these briefly yesterday, but if I
jdy> understand what Oleg is saying, the promise will only get evaluated when
jdy> it is needed, which might be later than the expression in which it
jdy> occurs.  The idea seems to be that passing promises around is cheaper
jdy> than matrices.
jdy> 
jdy> Perhaps the expression template could build, under some circumstances,
jdy> a promise rather than an actual result?
jdy> 
jdy> Joel
jdy> 
jdy> > On Tue, 20 Mar 2001, Joel Young wrote:
jdy> > jdy> lazy matrices and promises?
jdy> > jdy> 
jdy> > jdy> http://www.lh.com/~oleg/ftp/c++-digest/more_promises.txt
jdy> 
jdy> List-Unsubscribe: <mailto:boost-unsubscribe_at_[hidden]>
jdy>  
jdy> 
jdy> Your use of Yahoo! Groups is subject to http://docs.yahoo.com/info/terms/ 
jdy> 
jdy> 
jdy> 

----------------------------------------------------------------------
 Jeremy Siek                        www: http://www.lsc.nd.edu/~jsiek/
 Ph.D. Candidate                  email: jsiek_at_[hidden]
 Univ. of Notre Dame         work phone: [phone]
----------------------------------------------------------------------