Re: 0.8.2 candidate
"Oleh Derevenko" <[email protected]>
| Newsgroups | gmane.comp.lib.ode |
|---|---|
| Message-ID | <[email protected]> |
OK. Then the only thing we need to do is to use dNormalize3 in collision_trimesh_trimesh_new.cpp instead of inlined normalization operations. I'll have time next week-end if nobody else would like to try that change and test how it affects the demo application. Oleh Derevenko -- ICQ: 36361783 ----- Original Message ----- From: "Remi Ricard" <[email protected]> To: "Oleh Derevenko" <[email protected]>; "ODE Mailing List (E-mail)" <[email protected]> Sent: 2 жовтня 2007 р. 19:12 Subject: Re: [ODE] 0.8.2 candidate > Hi Oleh, > >> Well, why don't we then add another variant of normalization function >> which would find maximal component and divide by it before length >> calculation? That function could be used everywhere where it is known >> that vectors could be rather long. >> > > This is not what it is done in function dNormalize3 from the file > odemath.cpp > > > Remi >> Oleh Derevenko >> -- ICQ: 36361783 >> >> ----- Original Message ----- >> From: "Jon Watte" <[email protected]> >> To: "Oleh Derevenko" >> >> >> >>> If you do the analysis of how the bit errors propagate, the distortion >>> in direction is insignificant compared to the length problem. >>> >>> >>> Oleh Derevenko wrote: >>> >>>> By introducing extra multiplication before length calculation you will >>>> decrease error of addition operation but introduce more distortion to >>>> vector >>>> direction because that multiplication will be a separate operation and >>>> its >>>> result will be truncated to precision of float as well. >>>> >>>> >> >> >> _______________________________________________ >> ODE mailing list >> [email protected] >> http://ode.org/mailman/listinfo/ode >> >> > > > _______________________________________________ ODE mailing list [email protected] http://ode.org/mailman/listinfo/ode