Physicsless Parts vs. Lag

[SlithyTove]

Trying to get a better understanding of part counts and lag specifically as it relates to physicsless parts.

It seems like physicsless parts would have vastly less lag added than a physicsful part and that seems validated from other forum posts I have seen. But I'm not sure how to quantify that when I am building spacecraft. Can I slap them on mostly "free"?

Also, it seems like struts would be a special case for lag since they meaningfully affect the physics of the parents? IE a strut would add more lag than, say, a battery, but less lag than a physicsful part?

Appreciate any input/experience.

Edited October 16, 2015 by SlithyTove
Flipping to answered

[Gaarst]

Physicsless part is an obsolete denomination.

Before 1.0, such parts didn't have mass, nor drag: they didn't influence at all the physics of your ship with the exception of the link between them and their parent part.

Since 1.0 this has been corrected, and now "physicsless parts" have a mass, and cause drag if exposed to airflow.

The "new" physicsless parts (i don't know if those are the same as old physicsless parts) have for only particularity that they add their mass to the parent part and therefore do not need to be placed in symmetry to keep balance (this might also be true for drag but I'm not sure).

So, I'd say that they add as much lag as other parts do, with no noticeable difference.

[Claw]

...they add their mass to the parent part and therefore do not need to be placed in symmetry to keep balance (this might also be true for drag but I'm not sure).

Yes, this is all true (drag and mass).

Cheers,

-Claw

[Red Iron Crown]

Depends on the part. If it is a simple structural part like a cubic octagonal strut then the performance impact is minimal (and less than physicsful parts because the part has no mass and inertia of its own, the parent part's is simply increased). If it is a resource-using part then a lot of them can cause lag problems, like if you use too many RCS thrusters.

[MartGonzo]

Physicsless part is an obsolete denomination.

Before 1.0, such parts didn't have mass, nor drag: they didn't influence at all the physics of your ship with the exception of the link between them and their parent part.

Since 1.0 this has been corrected, and now "physicsless parts" have a mass, and cause drag if exposed to airflow.

The "new" physicsless parts (i don't know if those are the same as old physicsless parts) have for only particularity that they add their mass to the parent part and therefore do not need to be placed in symmetry to keep balance (this might also be true for drag but I'm not sure).

So, I'd say that they add as much lag as other parts do, with no noticeable difference.

I didn't realise that, interesting

I'm curious how offsetting a part into the body of a rocket affects drag, if at all?

anyone know?

[Gaarst]

I think I read recently that clipping a physicsless part inside the parent part does not reduce its drag. Don't quote me on that.

[MartGonzo]

I think I read recently that clipping a physicsless part inside the parent part does not reduce its drag. Don't quote me on that.

Thanks, I assumed that was the case, but thought I'd ask.

[Red Iron Crown]

It's more general than that, clipping *any* part does not affect its drag unless it is clipped into a cargo bay or service bay.

[Lynch]

Struts decrease the FPS not lag, while increasing the part count. The FPS decrease is due to calculation of less physical bodies as they are now connected to each other. However there is also a limit to where the decrease can happen. After using overdose of struts the game does start to fail.

Most advanced structures brough up need you to "Slap them on mostly" unless you use some special mods.

[MartGonzo]

It's more general than that, clipping *any* part does not affect its drag unless it is clipped into a cargo bay or service bay.

good to know, thanks.

[Snark]

Incidentally, if you're ever wondering "what effect on aero forces does part X have if placed in situation Y", you can toggle on the aero overlay via F12 (though you'll need to muzzle Steam to do that, since it uses F12 as a hotkey and gets in the way). Adds a bunch of informative arrows to the graphic display, whose length and direction shows you what's going on.

It's not perfect (e.g. if you have a part clipped inside something, chances are good that you won't see its aero arrow because the arrow may also be inside the part), but it can be interestingly informative.

[SlithyTove]

Physicsless part is an obsolete denomination.

Before 1.0, such parts didn't have mass, nor drag: they didn't influence at all the physics of your ship with the exception of the link between them and their parent part.

Since 1.0 this has been corrected, and now "physicsless parts" have a mass, and cause drag if exposed to airflow.

The "new" physicsless parts (i don't know if those are the same as old physicsless parts) have for only particularity that they add their mass to the parent part and therefore do not need to be placed in symmetry to keep balance (this might also be true for drag but I'm not sure).

So, I'd say that they add as much lag as other parts do, with no noticeable difference.

Thanks for some of the clarifications!

Still, seems like under the new model the physicsless parts would still require less overall computation to keep running since they can't move themselves independent of their attached body and doing on calculation with added drag and mass on the main part is still quicker than a calculation on two independent parts.

Or am I missing something?

[SlithyTove]

Well, I just did a trial run with a simple rocket stack with 1000 of the smallest batteries attached and compared that to the same rocket stack with 1000 of the small structural panels attached.

If there is a difference in performance between the two ships it's subtle. Both were running about 8 fps with 3.5-4s of real time elapsed per game second.

That is with all graphical settings turned down to the minimum.

That's a bit surprising to me, as I can see in the aero overlay that each one of those structural panels has it's own aeroforce indicator, but the battery stack only had aero indicators for the main stack.

In that case at least, a part is a part is a part more or less.

[Claw]

The only difference in the test that you ran is where the final force is applied. All of the drag calculations work the same, but the final product is passed up to the parent to be applied, vice being applied at the physicsless part. Being physicsless doesn't really save or cost any different than a "regular" part.

Where you will notice a difference is if you attach several physicsless parts to only one side of a craft (so it's asymmetric) and fly it. Then do the same with a bunch of small "regular" parts (again, asymmetric). In this case, you'll see a difference in the stability of the craft. The positioning of non-physicless parts on a parent will affect stability a lot more than for physicsless parts, which is really the main purpose now. (Both for mass and drag.) i.e. so you can put a ladder on a small ship, or a science part on a small probe without making it annoyingly difficult to control.

Cheers,

-Claw

Edited October 15, 2015 by Claw

[SlithyTove]

Thanks Claw!

Every post I have seen from you has been impressively informative and helpful, and this is no different.

Appreciate all the other input, going to drop this one to answered.