Grasshopper 0.7.0014 available for download

Replies to This Discussion

Permalink Reply by David Rutten on May 20, 2010 at 4:33am

Hi Enrique,

although it is true that many algorithms can be described in Big O notation, it is very difficult to figure out the runtime from the outside. A lot of functions in Rhino (and Grasshopper) have optimizations build in which might only kick in for certain cases.

But even if I managed to somehow display a representation of the runtime efficiency of different components, would it really help you? I mean, it is unlikely that you are computing irrelevant data, so whatever operations you are performing have to be performed in order to reach the desired result.

The current profiler widget goes some way towards providing some insight into the relative speeds of different objects (btw the current implementation is not to be trusted, I was incorrectly measuring the same cycles for multiple components in some cases).

--
David Rutten
david@mcneel.com
London, UK

Permalink Reply by quitronic on May 18, 2010 at 8:26pm

David:

In advance, my apologies for my bad English, it isn’t my first language.
In another more complex definition, i have found a bug with the Orient component.
Here, a simpler definition of the problem.

I need the tree structure for the problem. The oriented geometry should be placed in the target XY Planes but not in another XY Planes.
Is a Orient component problem? or, my problem?
I hope it is understood.
thanks in advance!

Attachments:

• Orient BUG.ghx, 78 KB

Permalink Reply by tangminghao on May 19, 2010 at 12:22am

Thank you very much!

Permalink Reply by Carlo Beltracchi on May 19, 2010 at 3:53am

That's great!

Thank you David!!!!

Permalink Reply by Tuan N. Tran on May 20, 2010 at 12:09am

some bugs to report:

1) Galapagos genome connections to sliders are lost if sliders and Galapagos components are copied then paste into another file.
2) Some math operators persistent data after copy/paste become NULL. i can remember ADD for sure.
3) Evaluate Curve automatically reparameterizes the Curve. Curve frame does the same.

wishes:

1) "quick" connect to all selected sliders to Galapagos
2) sliders have option to connect/disconnect from Galapagos. problematic with multiple Galapagos components?

thanks again for a great release.

Permalink Reply by David Rutten on May 20, 2010 at 4:28am

Thanks Tuan,

I'll try to fix this (except for the Galapagos bug, not critical enough) for the next release.

--
David Rutten
david@mcneel.com
London, UK

Permalink Reply by Tuan N. Tran on May 20, 2010 at 7:28am

here's another one for you: curve shatter produces an "Invalid Curve" with multiple t-param inputs.

Permalink Reply by WT4 on May 20, 2010 at 9:17am

Hi David

In the hillclimber example you are finding the maximum height whereas in the catenary example
the galapagos component is looking for the minimum distance. How does it know whether to look
for the maximum or minimum, ie if the hill terrain is turned upside down and the hills are now
ravines. How do you then tell the galapagos component to look for the minimum height.

Many thanks
William

Permalink Reply by Ángel Linares on May 20, 2010 at 9:33am

Good question I'll keep reading...¬¬

Permalink Reply by Danny Boyes on May 20, 2010 at 9:35am

I'd like to know the right way of doing this too. As a work around I'm looking for the inverse ie dividing one by my distance.... is this right?

Permalink Reply by Vicente Soler on May 22, 2010 at 3:33pm

Galapagos always tries to find a larger number as the result of the fitness function. For the catenary you would use 1/distance. The smaller the distance the larger is the resulting number of the fitness function.

To reverse the hills into ravines my guess is to just multiply the z coordinate by -1 but i haven't tried it.

Permalink Reply by WT4 on May 20, 2010 at 10:53am

In the enclosed image, I am trying to maximise the ratio of the volume/surface area of a box. Can anyone
explain the scatter and line graphs that appear. Also what is the difference between the dots and the cross markers.

Also if I want to use this galapagos component as an optimisation tool, where the objective is to minimise
say the weight of a beam component of a fixed length. Clearly the minimum weight can be obtained by
choosing the smallest cross-section. This clearly will not satisfy other requirement such as the maximum stress allowable. Is it therefore possible to incorporate design constraints into the solution

Regards
William

• ‹ Previous
• 1
• 2
• 3
• 4
• 5
• 6
• Next ›
• Page