Internalasing couple ten millions of numbers freeze

Replies to This Discussion

Permalink Reply by Mateusz Zwierzycki on July 7, 2016 at 3:46pm

In this case for each number the internalization function is called separately... Be them stored in one list or in a data tree. Basically a couple tens of millions is a lot, no matter what program you're using... deserializing will also take some time, are you sure you want to internalize them rather than load from the source file ?

EDIT: Oh, you have them from some other component which gives you the daylight results... then you need some script/plugin to save them to a text file and load it from a text file afterwards. I guess displaying those numbers in the panel might also crash grasshopper. 

Permalink Reply by Jonas Blazinskas on July 7, 2016 at 3:58pm

I think I can reduce it slightly by path mapping it would then have aroun 600 numbers in each list, if that doesn't work then yeah i need to find away to export

Permalink Reply by Jonas Blazinskas on July 7, 2016 at 3:59pm

oops I didin't reply but edited the message somehow  

Permalink Reply by Jonas Blazinskas on July 7, 2016 at 4:47pm

I don't get it one simulation takes 2 minutes to produce all the results internalising one takes few seconds. Saving for an infinite time? Something does't add up here

Permalink Reply by Jonas Blazinskas on July 7, 2016 at 5:51pm

3,506 KB file taking 5 minutes to save and brb going for a smoke because this is stressing me out we will see how long does it take to open it. Its just 1s and 0s in 10k branches do branches lag so much?

Permalink Reply by Jonas Blazinskas on July 7, 2016 at 5:53pm

Something has to be bugged why doesn't it have the widget either? 

Permalink Reply by David Rutten on July 7, 2016 at 11:53pm

I'll need to look into all this some more, but the profiler only counts time spent on the solution, it doesn't account for time spent doing other things such as deserialising.

Permalink Reply by Jonas Blazinskas on July 8, 2016 at 12:00am

I suspect a few things that could have went wrong on my part. They used to be booleans and I attached them to num component which adds more complexity i bet by making the number go from 1 to 1.0 maybe something on the transfer bugs out the results as well. I believe that the amount of branches played a huge role for the lag there as well. 

So why does the top num component has the profiler? all 3 of these components only have internalised data