Hello, I have almost 2 solutions to implements a module. I there a way to test performance (time and/or nb triangles) of each implementation ? "F6 render" show some informations (cleaning cache every time) but my module is too small to display something useful. I try to duplicate the module with a loop on children but I have the same result in two case so I suspect that I measure duplication time and not my module time ! One module use polygon with point list. Second one use difference(). a+ Vicnet -- Sent from: http://forum.openscad.org/

I don't understand what the question is. When you preview or render, it reports the number of seconds it took to process the geometry. If the number is too high for your taste, then you can try to make more efficient module etc. and see if it runs in less time. On Thu, Mar 22, 2018 at 8:03 AM, vicnet <vo.publique@gmail.com> wrote: > Hello, > > I have almost 2 solutions to implements a module. > I there a way to test performance (time and/or nb triangles) of each > implementation ? > > "F6 render" show some informations (cleaning cache every time) but my module > is too small to display something useful. > I try to duplicate the module with a loop on children but I have the same > result in two case so I suspect that I measure duplication time and not my > module time ! > > > One module use polygon with point list. > Second one use difference(). > > a+ > Vicnet > > > > -- > Sent from: http://forum.openscad.org/ > > _______________________________________________ > OpenSCAD mailing list > Discuss@lists.openscad.org > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org

Em 22 de mar de 2018 13:42, "Hans L" <thehans@gmail.com> escreveu: > I don't understand what the question is. When you preview or render, > it reports the number of seconds it took to process the geometry. If > the number is too high for your taste, then you can try to make more > efficient module etc. and see if it runs in less time. > > > On Thu, Mar 22, 2018 at 8:03 AM, vicnet <vo.publique@gmail.com> wrote: > > Hello, > > > > I have almost 2 solutions to implements a module. > > I there a way to test performance (time and/or nb triangles) of each > > implementation ? > > > > "F6 render" show some informations (cleaning cache every time) but my > module > > is too small to display something useful. > > I try to duplicate the module with a loop on children but I have the same > > result in two case so I suspect that I measure duplication time and not > my > > module time ! > > > > > > One module use polygon with point list. > > Second one use difference(). > > > > a+ > > Vicnet > > > > > > > > -- > > Sent from: http://forum.openscad.org/ > > > > _______________________________________________ > > OpenSCAD mailing list > > Discuss@lists.openscad.org > > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org > > _______________________________________________ > OpenSCAD mailing list > Discuss@lists.openscad.org > http://lists.openscad.org/mailman/listinfo/discuss_lists.openscad.org >

> I try to duplicate the module with a loop on children but I have the same > result in two case so I suspect that I measure duplication time and not my > module time ! Have you tried not using children (which I suppose is just making copies of the already generated object) and instead repeatedly call the module in a loop? -- Sent from: http://forum.openscad.org/

> > One module use polygon with point list. > Second one use difference(). > ​I would not expect big differences with 2D models. Anyway, remember that OpenSCAD saves models in cache so there is no duplicate effort if the very same model is used many times.​

> I don't understand what the question is. When you preview or render, > it reports the number of seconds it took to process the geometry. If > the number is too high for your taste, then you can try to make more > efficient module etc. and see if it runs in less time. The render value is fine to have some perf values. The pb is how to test 2 implementations that take few time (milliseconds o rless?) each so that it is not comparable for one call with render? Troberg wrote > Have you tried not using children (which I suppose is just making copies > of > the already generated object) and instead repeatedly call the module in a > loop? Yes, you're right. i use children() to duplicate so certainly cache is used and I measure children+cache functions not my module ! I will try to duplicate directly. Thanks. a+ vicnet -- Sent from: http://forum.openscad.org/