Alternatives to hoopsnake?

Hi!

For a new random test I need something to avoid recursive data errors...

Hoopsnake should be good, but the problem for me is that it STORE every cycle in its memory!

The first cycles are fast, but then it become to slow everything very shortly, because it is created a huge tree of datas...

There is something else very very (VERY) simple, to just avoid recursive data errors?

just like:

data>to memory

memory>to new data

without any other things, just... as light as possible....

anything?

maybe by c#?

thanks in advance

Replies to This Discussion

Permalink Reply by ng5 Alex on January 16, 2015 at 7:25am

Hello

have you tried anemone?

dont know how ti handles data, but maybe its suitable for your need.

Permalink Reply by Daniel Piker on January 16, 2015 at 7:36am

The loop component in Octopus seems to work well: http://www.grasshopper3d.com/m/photo?id=2985220%3APhoto%3A1122752

Permalink Reply by Riccardo Majewski on January 16, 2015 at 7:56am

Thanks guys for fast reply!

I didn't know either of them.

I'm trying them now.

Thanks again!

Permalink Reply by Mateusz Zwierzycki on January 16, 2015 at 8:11am

It may also be that you don't manage the data paths correctly and in the end they become very long (each loop appends new digits to the datatree). When those paths become very long (like 100 digits), it slows down the definition tremendously.

Permalink Reply by Riccardo Majewski on January 16, 2015 at 8:23am

I don't really remember about data paths in the output;

I just know that I really don't need a history of older datas, those are completely useless for me, as I'm trying to implement it in a simulation definition.

Hoopsnake have a slot where it store ALL old datas, and there is no way to disable this, thus weighing for nothig.

Also anemone is slow:

just in the tutorial where for 10 very simple loops (like x=x+1) it takes fractions of a second, where it should be almost instant in every cpu by now

Permalink Reply by Mateusz Zwierzycki on January 16, 2015 at 9:10am

Also anemone is slow:

just in the tutorial where for 10 very simple loops (like x=x+1) it takes fractions of a second, where it should be almost instant in every cpu by now

working on that...

Permalink Reply by peter fotiadis on January 16, 2015 at 9:17am

Describe exactly what you are after and a C# script could be near you in no time.

Permalink Reply by Riccardo Majewski on January 16, 2015 at 9:41am

Working with kangaroo, I'm trying to change vector trajectories depending of relatives positions and angles of vectors.

This should be done instant by instant.

Hoopsnake could be good but it grows an infinite pile of useless data, as said, weighting all for nothing.

Anemone can have history disabled but i't apparently slow, for me.

Still I haven't managed to get Octopus work.

What I SUPPOSE could be done is similar to this:

(and I don't know almost anything in C# and similar)

IF c# script restart, means input value changed (?)

(but this is not a really needed point, as in a simulation like mine, every iteration is different)

- take data from GH and send to specific located "slot" (or whatever) in memory (RAM uh?)

- take data back from RAM and output to GH

1 pin input 1 pin output

Hoopsnake really is close to this, just the "cumulative Feedback" ruins everything.

Permalink Reply by peter fotiadis on January 16, 2015 at 10:04am

Er... let's put out RAM from out discussion, he he.

Let's roll it again (it's really elementary with C#):

You mean that you want to send some collection (DataTree and/or List) of random things (Vectors I guess) to Kangaroo?

If yes provide some indicative static definition of yours that does what you want (gets some input and makes some output). By static I mean that it does something once (and/or without taking advantage of other factors that could affect the "randomness" and/or the whole iteration logic).

Permalink Reply by Riccardo Majewski on January 16, 2015 at 10:46am

Thanks for the help!

Ok for the RAM part XD

I have no idea of how make an example of it in only 1 step/iteration...

It's a swarming model, roughly...

Permalink Reply by peter fotiadis on January 16, 2015 at 10:59am

Hmm ... a sketch maybe ? a 123 page description?

I mean you said :

I'm trying to change vector trajectories depending of relatives positions and angles of vectors.

This unfortunately can't describe what the C# should do.

Something "swarming" like these maybe?

Attachments: