Re: surface cdata property restricted to double or uint8

[Jaroslav Hajek]

On Mon, Aug 30, 2010 at 11:24 AM, Michael Goffioul
<address@hidden> wrote:
> On Mon, Aug 30, 2010 at 10:02 AM, Jaroslav Hajek <address@hidden> wrote:
>> Whether or not an array is temporary is not that much relevant; the
>> important point is that the array is hanging around. If I'm working
>> with arrays of 10^7 doubles, saving a temp of 10^6 doubles is
>> marginal, even though it may be a huge saving otherwise. But you're
>> indeed right that cmapv is not 3*lda long. So put 25% there.
>
> It's not a factor 10, but more a factor 3, which means you can gain about
> 30-33% memory. If you can do it, why not?
>

Percentage is normally stated relative to status quo. You'd cut down
memory from 4*lda to 3*lda, that is -25%. Still not much (considering
that Octave is not really memory effective), but perhaps already worth
it. My earlier suggestion was based on my incorrect 14% estimate.


-- 
RNDr. Jaroslav Hajek, PhD
computing expert & GNU Octave developer
Aeronautical Research and Test Institute (VZLU)
Prague, Czech Republic
url: www.highegg.matfyz.cz