On Sunday 01 December 2002 13:38, mth wrote:
> >> Q: What was the reasoning behind using doubles instead of floats? Is
> >> it the case that this level of precision (or range) is needed?
>
> Is there any type of calculation in chemistry where it would ever make
> sense to have that much range & precision?
I guess Quantum Mechanics...
> > I did not report it yet, as I was wondering if the increased precision
> > had effects on the performance... (which it should...)
>
> At one level, some of the biggest challenges in improving performance are
> bus bandwidth and cache hit ration. So moving around 8 bytes at a time is
> more expensive than moving around 4 bytes at a time. Because it uses more
> bus bandwidth and reduces cache effectiveness.
>
> > but haven't had
> > time to do some real test... though the effect seems rather marginal...
>
> I am sure that the direct effects are marginal.
> But, in my opinion, Unless there is a real application-level need to do
> it, using 8 bytes when 4 will do is wasteful.
>
> We could use 8 byte integers for everything too, but we don't. In most
> cases there is no need for it.
True.
> > but this may also be caused by the removal of a number of method
> > calls... which was one of the advantageous effects of the move to CDK
> > ...
>
> I assume that moving to CDK is a good thing.
Yes, it is ;)
The advantages:
1. CDK has a larger developers community
2. CDK has a large repository of chemoinformatical libraries
(substructure search, download from internet, database backend,
2D viewer, etc)
3. CDK has IO to JOELib and will hav IO to CMLDOM in the future
4. CDK will enlarge Jmol's user base
Moreover,
5. CDK was developed to replace the current Jmol Atom, Bond classes
BTW, replacing Jmol's classes is going reasonably smooth... at least, it going
easier than I originally thought...
Egon
Egon
-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf
_______________________________________________
Jmol-developers mailing list
[EMAIL PROTECTED]
https://lists.sourceforge.net/lists/listinfo/jmol-developers
