Build 293.10 - Release version

Quote:

---

Originally Posted by defisym

I'm now using the steam beta version, R293.8. After a bit research I managed to reproduce it:
1. create a new "Test" mfa, insert an active to frame.
2. drag an active ("char-1001") from another "Data" mfa(with different props, e.g. DX9 vs DX11)
3. export animation from the object dragged in ("char-1001"), and load it by the first active, then the issue can be reproduced
4. close and reopen the "Test" mfa can fix this issue. That might because some snyc issue when dragging objects between mfas

---

Hmm, interesting, I'll fix it.

I just noticed a few options descriptions that I never noticed before, these may be new...

Optimize events: May modify events to favorize speed.
Made me wonder how these optimized events would look, could we could learn from it.

So, would be possible to have this option as an feature that you call from inside event editor, where it modify all the events in a way we can actually see?
(May also help me suggesting more optimizations)

Quote:

---

Originally Posted by NaitorStudios

I just noticed a few options descriptions that I never noticed before, these may be new...

Optimize events: May modify events to favorize speed.
Made me wonder how these optimized events would look, could we could learn from it.

So, would be possible to have this option as an feature that you call from inside event editor, where it modify all the events in a way we can actually see?
(May also help me suggesting more optimizations)

---

Yes and no. It does 2 types of modifications:

1. It re-orders conditions to make sure faster conditions are executed first. As very simple example, if you do:

- Object A overlaps Object B
+ Global value A = 1

It will exchange both conditions as testing a global value is much faster than testing collisions and the other condition won't be evaluated if Global Value A is different from 1. This could be potentially added as a manual option for each event, but well...

2. It combines some actions and conditions that can be executed together. This can't be done in the editor.

Ah I see, I thought it would do things a bit differently, like merging events with same conditions.
I guess Optimize events also take care of not moving conditions in a way that would break scoping, right?

One thing that I noticed is that a flag can be more efficient than using break, and using a flag to keep the overlap state instead of checking on multiple events is also a lot better.
Although it shouldn't be done in events that actually moves the object that you need to test for overlap.

I'm planning on making a tool that auto optimizes events with child events and some other stuff.
(it would read and interpret the events in form of text that we can copy to clipboard with 2.5+, the rest would be just an automation macro to copy and paste stuff around)

Could we get a condition like "Pick all objects in a line" for specific objects, like Active and Qualifiers?
And perhaps a expression that outputs the amount of selected objects... Please?
"Pick all objects in a line" from "New Objects" selects anything, would conflict with the expression above...

Was trying to avoid using extensions like Select Object or Scope Control because of these are Windows only, and in thing I'm working on, running a For Each loop isn't doing it like I was expecting.

Quote:

---

Originally Posted by NaitorStudios

I guess Optimize events also take care of not moving conditions in a way that would break scoping, right?

---

Correct, it won't optimize if there is a risk to break scoping.

Quote:

---

One thing that I noticed is that a flag can be more efficient than using break.

---

I'm curious to see in which case this is more efficient.

Quote:

---

Originally Posted by Yves

I'm curious to see in which case this is more efficient.

---

Here's a example, you can see that with the flag it got a lot more FPS than using Break. 23% more efficient with a Flag. (44,4% more efficient than Negated Overlap)

Attachment 30700

Crazy how much the Debugger affects performance with that example
Debugger OFF 1000fps Debugger ON 500fps
A ram usage expression in fusion would be good for custom debuggers

Quote:

---

Originally Posted by NaitorStudios

Here's a example, you can see that with the flag it got a lot more FPS than using Break. 23% more efficient with a Flag. (44,4% more efficient than Negated Overlap)

Attachment 30700

---

Thanks, I'll need to check in depth what happens in your example, because this doesn't seem normal (and on another machine because on mine it's a steady 1000 fps in both cases).

Update: the build 293.9 is available, hopefully next official version.