Poll

What do you like doing most in RA2/RA3/robot combat games?

Building robots
Local single player battles with manually-controlled robots
Local multiplayer battles  with manually-controlled robots (PvP with controllers or a shared keyboard)
AI-only tournament battles
Other?  Please comment below.

Author Topic: Robot Rumble 2.0 - Robot Combat Simulator - Under Development  (Read 187602 times)

Offline kix

  • RR2 dev
  • *
  • Posts: 3452
  • Rep: -3
  • H
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1760 on: February 13, 2020, 12:03:16 PM »
How's your CPU/GPU utilisation while playing? No change in FPS from 768p to 4k is a good pointer to it being a huge CPU bottleneck, possibly due to your CPU not being able to keep up with the physics engine? Try lowering the physics engine tickrate, if that's an option in the game. If neither your GPU or CPU are pinned at 100% utilisation, I'd take a wild uneducated guess at there being a weird bottleneck somewhere in the rending pipeline.
The game has tickreate
however 100tick is not desirable due to wonky driving, aka bots bounce while driving.
My cpu/gpu never reach 50% in the game at all, no matter what tick

Offline CodeSilver23

  • RR2 Enthusiast and Design Innovator
  • Heavyweight
  • Posts: 561
  • Rep: -2
  • Stars v Legends founder, Inventor of the Wheel-fix
    • https://www.youtube.com/c
    • View Profile
    • Youtube Channel
    • Awards
  • See profile for gamer tags: Yes
  • Discord: CodeSilverGaming
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1761 on: February 13, 2020, 12:24:45 PM »
I think I might have come up with a potential solution for the motor gyro phenomena. Won’t be able to test it until tonight, so I’ll let you guys know then.
Accomplishments:

Offline CodeSilver23

  • RR2 Enthusiast and Design Innovator
  • Heavyweight
  • Posts: 561
  • Rep: -2
  • Stars v Legends founder, Inventor of the Wheel-fix
    • https://www.youtube.com/c
    • View Profile
    • Youtube Channel
    • Awards
  • See profile for gamer tags: Yes
  • Discord: CodeSilverGaming
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1762 on: February 13, 2020, 06:47:52 PM »
Also, for those who don't know what I'm talking about, here is a basic visual:


My idea didn't work, so cjbruce please fix this.
Accomplishments:

Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1763 on: February 14, 2020, 06:39:49 AM »
Sorry for being incommunicado.  I’ve been out with the flu for the past few days.

CodeSilver, thank you for the video.  It is super clear and easy to reproduce.   When I get my strength back I will have a look.

kix, by any chance are you using KupaTech’s screen resolution hack?  The way I understand it, the hack sets the resolution of the screen one frame after the screen loads.  This means you will never see any resolution other than the one hard-coded by the hack.

Offline kix

  • RR2 dev
  • *
  • Posts: 3452
  • Rep: -3
  • H
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1764 on: February 14, 2020, 07:07:05 AM »
kix, by any chance are you using KupaTech’s screen resolution hack?  The way I understand it, the hack sets the resolution of the screen one frame after the screen loads.  This means you will never see any resolution other than the one hard-coded by the hack.
Uh i do, however you can turn it off which is convenient. I turned it off when i did the testing

Offline Team Lightning

  • Addicted to combat robotics, good at annoying people, and wanting to get to -10000 rep XD.
  • Beetleweight
  • Posts: 168
  • Rep: -4
  • Education is important, robots are importanter.
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1765 on: February 14, 2020, 04:37:39 PM »
Sorry for being incommunicado.  I’ve been out with the flu for the past few days.

CodeSilver, thank you for the video.  It is super clear and easy to reproduce.   When I get my strength back I will have a look.

kix, by any chance are you using KupaTech’s screen resolution hack?  The way I understand it, the hack sets the resolution of the screen one frame after the screen loads.  This means you will never see any resolution other than the one hard-coded by the hack.
Can you post the code for this hack?

Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1766 on: February 14, 2020, 04:59:19 PM »
Also, for those who don't know what I'm talking about, here is a basic visual:


My idea didn't work, so cjbruce please fix this.

CodeSilver23, you are a genius.  With the help of the video I was able to find an eliminate the problem.  The errant gyroscopic effect that you have shown should be gone in the next build.

For the physics geeks out there:

Unity has really lousy physics documentation.  I had assumed that when you change the mass of an attached rigidbody, the moments of inertia would be recomputed as well.  In our case, I was setting the mass of the rigidbody to a very small number (0.000001 kg) when all of the components break off an axle.  I had assumed (incorrectly) that this would cause the moments of inertia (which directly affect angular momentum and gyroscopic effects) to be nearly zero as well.

I was wrong.  CodeSilver23 was spot on in his video.  What he was seeing could only be gyroscopic effects.  I turned on a display of moment of inertia and sure enough, the MOI was set to 1 kg * m^2 in all dimensions.  This is an insanely huge moment of inertia for a wheel.  Apparently, Unity decided that since it could no longer compute an MOI because there was nothing attached, instead of setting the MOI to 0 it set MOI to 1.  This is crazy.  It also explains the ridiculousness that happens when you break everything off.

In any case, it should be fixed now.  I tried to go back and find and fix all of the weird quirks that I introduced when I didn't know the root cause of the problem.  Hopefully I got everything, but please let me know if you find more stuff cropping up in the next build.

Offline kix

  • RR2 dev
  • *
  • Posts: 3452
  • Rep: -3
  • H
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1767 on: February 14, 2020, 05:09:46 PM »
Also, for those who don't know what I'm talking about, here is a basic visual:


My idea didn't work, so cjbruce please fix this.

CodeSilver23, you are a genius.  With the help of the video I was able to find an eliminate the problem.  The errant gyroscopic effect that you have shown should be gone in the next build.

For the physics geeks out there:

Unity has really lousy physics documentation.  I had assumed that when you change the mass of an attached rigidbody, the moments of inertia would be recomputed as well.  In our case, I was setting the mass of the rigidbody to a very small number (0.000001 kg) when all of the components break off an axle.  I had assumed (incorrectly) that this would cause the moments of inertia (which directly affect angular momentum and gyroscopic effects) to be nearly zero as well.

I was wrong.  CodeSilver23 was spot on in his video.  What he was seeing could only be gyroscopic effects.  I turned on a display of moment of inertia and sure enough, the MOI was set to 1 kg * m^2 in all dimensions.  This is an insanely huge moment of inertia for a wheel.  Apparently, Unity decided that since it could no longer compute an MOI because there was nothing attached, instead of setting the MOI to 0 it set MOI to 1.  This is crazy.  It also explains the ridiculousness that happens when you break everything off.

In any case, it should be fixed now.  I tried to go back and find and fix all of the weird quirks that I introduced when I didn't know the root cause of the problem.  Hopefully I got everything, but please let me know if you find more stuff cropping up in the next build.

Thats why we are here, to test stuff and troubleshoot them.

Now how about the wheels jiggling when they are not touching the ground at 100tick?

Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1768 on: February 14, 2020, 05:23:21 PM »
Ok so i did some testing around with the bots that have 600+ colliders, thanks rep makers, appreciate that.


One thing im having a feeling about is that the game may be only limiting itself to max 4 cores/ 4 threads. Im not sure on this one, maybe thats why my cpu usage cant reach 100% (6c/12t), again, im not exactly sure on this one.

To be honest, I'm not a skilled enough programmer to figure out how to split the biggest CPU loads over more than a single core.  Physics is the biggest load right now, and it is notoriously difficult and/or impossible to split it over multiple CPUs.

And uh ok so resolution settings do nothing, nothing at all, no fps gain/decrease.

Even did windowed mode of the game and scaled down the window down to this

even changed the res to 4k as in the yt vid

Would it be possible for you to run the plain-vanilla version of the game without the KupaTec hack?  On my end the screen resolution settings work.  The resizing/scaling is wonky, but KupaTec is actually planning to do a rewrite of this for the near future.

On my end, the more I boost up the resolution, the more the frame rate drops.  Graphics quality doesn't have nearly as big of an effect as screen resolution, indicating my poor laptop GPUs are definitely fill rate limited.



Then i tried gfx quality settings.
Mid/ High are the same, low however


It might be shadows or post processing that is making stuff lag.


Its hard to say.  It is surprising that your computer has issues with graphics settings on low.  What kind of a graphics card do you have?


Also what ive noticed, when the bots are upside down, and if you drive the motors at full speed, on 100 tick, the premade wheels wheels spaz out.
EDIT: Premade wheels do that too:

Let me look into this.  I suspect the crazy high RPMs are throwing off the wheels.  It might not be easily fixable.  Let me think about it a little bit.

Offline kix

  • RR2 dev
  • *
  • Posts: 3452
  • Rep: -3
  • H
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1769 on: February 14, 2020, 05:51:47 PM »
Ok so:
A. I could try and yeah it would change the fps. However even in vanilla version pre kupatec thing would not really change the fps. The difference is only around 100fps. (And the game is doing 200+fps on highest settings)

B. The game does not have issues on low. You see its opposite. Fps gets better, however what i noticed on low is that the shadows are off, and maybe each part has its own shadows to calculate which kills the fps on high part count bots. My gpu is a 8gb Sapphire RX580

C. Weird thing is that light parts on a spinner that are approx the same weight dont do this thing at all

Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1770 on: February 15, 2020, 08:51:16 AM »
I'm going to try to get out a build while I have a clear head.  This one has several things that change the game in major ways:

1. The chassis plates can now be broken off.  This means that every component that you add to the robot can become exposed.  This means that "Robot Health" effectively becomes a meaningless idea.  You can continue bashing pieces off a robot until there is nothing left to bash.  I'm strongly leaning toward eliminating the health bars at the top of the screen, as they don't have any meaning anymore.

The "Robot" itself is effectively the bottom plate of the chassis.  It can't be broken off from itself.  The camera tracks it.  The "P1" icon hovers above it.  All other pieces are attached to it via a tree structure.  One question is lingering in my mind: should the bottom chassis plate be invulnerable?  If not, what should the consequences be for hitting it?  On a good hit should we roll a die and randomly destroy something attached to it?


2. The first of the new arenas built with KupaTec's new "Arena Modding Tool" (AMT) is going to be included in this build.  It is a shameless duplicate of the school arena that my students will be competing at on February 28-29.  It is a smaller box, suitable for lightweights.  It has polycarbonate walls, 8" steel angle iron guard rails, and an arena spinner hazard.

The Arena Modding Tool makes it possible for you guys to create and share your own arenas.  KupaTec and WhamettNuht are the geniuses behind the tool and the new arenas.  Expect more arenas soon!


3. We are in the process of removing old arenas, robots, menu systems, UI, and other stuff that has developed and mutated over the years.  Over the next few months, expect to see a lot of stuff disappear and be replaced by newer, shinier versions.  And you know what that means? New bugs!!! 

All kidding aside, we are consolidating things down so that there is only one menu system (instead of one for every screen), one set of rules for robot behavior (instead of one for each robot), one set of rules for arena construction (instead of one for each arena), etc.  It should significantly reduce the amount of bugs that need to be squashed.


With this build you are seeing the start of the transition of the game from its current Alpha state into a future Beta state.

Offline Team Lightning

  • Addicted to combat robotics, good at annoying people, and wanting to get to -10000 rep XD.
  • Beetleweight
  • Posts: 168
  • Rep: -4
  • Education is important, robots are importanter.
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1771 on: February 15, 2020, 08:52:59 AM »
Also, for those who don't know what I'm talking about, here is a basic visual:


My idea didn't work, so cjbruce please fix this.

CodeSilver23, you are a genius.  With the help of the video I was able to find an eliminate the problem.  The errant gyroscopic effect that you have shown should be gone in the next build.

For the physics geeks out there:

Unity has really lousy physics documentation.  I had assumed that when you change the mass of an attached rigidbody, the moments of inertia would be recomputed as well.  In our case, I was setting the mass of the rigidbody to a very small number (0.000001 kg) when all of the components break off an axle.  I had assumed (incorrectly) that this would cause the moments of inertia (which directly affect angular momentum and gyroscopic effects) to be nearly zero as well.

I was wrong.  CodeSilver23 was spot on in his video.  What he was seeing could only be gyroscopic effects.  I turned on a display of moment of inertia and sure enough, the MOI was set to 1 kg * m^2 in all dimensions.  This is an insanely huge moment of inertia for a wheel.  Apparently, Unity decided that since it could no longer compute an MOI because there was nothing attached, instead of setting the MOI to 0 it set MOI to 1.  This is crazy.  It also explains the ridiculousness that happens when you break everything off.

In any case, it should be fixed now.  I tried to go back and find and fix all of the weird quirks that I introduced when I didn't know the root cause of the problem.  Hopefully I got everything, but please let me know if you find more stuff cropping up in the next build.
I'm going to try to get out a build while I have a clear head.  This one has several things that change the game in major ways:

1. The chassis plates can now be broken off.  This means that every component that you add to the robot can become exposed.  This means that "Robot Health" effectively becomes a meaningless idea.  You can continue bashing pieces off a robot until there is nothing left to bash.  I'm strongly leaning toward eliminating the health bars at the top of the screen, as they don't have any meaning anymore.

The "Robot" itself is effectively the bottom plate of the chassis.  It can't be broken off from itself.  The camera tracks it.  The "P1" icon hovers above it.  All other pieces are attached to it via a tree structure.  One question is lingering in my mind: should the bottom chassis plate be invulnerable?  If not, what should the consequences be for hitting it?  On a good hit should we roll a die and randomly destroy something attached to it?


2. The first of the new arenas built with KupaTec's new "Arena Modding Tool" (AMT) is going to be included in this build.  It is a shameless duplicate of the school arena that my students will be competing at on February 28-29.  It is a smaller box, suitable for lightweights.  It has polycarbonate walls, 8" steel angle iron guard rails, and an arena spinner hazard.

The Arena Modding Tool makes it possible for you guys to create and share your own arenas.  KupaTec and WhamettNuht are the geniuses behind the tool and the new arenas.  Expect more arenas soon!


3. We are in the process of removing old arenas, robots, menu systems, UI, and other stuff that has developed and mutated over the years.  Over the next few months, expect to see a lot of stuff disappear and be replaced by newer, shinier versions.  And you know what that means? New bugs!!! 

All kidding aside, we are consolidating things down so that there is only one menu system (instead of one for every screen), one set of rules for robot behavior (instead of one for each robot), one set of rules for arena construction (instead of one for each arena), etc.  It should significantly reduce the amount of bugs that need to be squashed.


With this build you are seeing the start of the transition of the game from its current Alpha state into a future Beta state.
When's this build coming out?
EDIT: I'm gonna guess soon?

Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1772 on: February 15, 2020, 09:09:29 AM »
The 15February2020 Alpha Build is out!

https://robot-rumble.itch.io/builds


It comes with a pretty big list of changes.  Thank you to everyone on the team and to everyone on GTM who contributed so much in the past few weeks! :claping

[Updates in the February 15th Build]​

[Added] Chassis armor Shape_Plates now pop off when their associated DamageableObject have health < 0.  This means that it is now possible to break off the protective armor plates of the chassis, exposing the motors on the inside.​

[Added] ARENA MODDING TOOL: Arenas created with the Arena Modding Tool are now playable.

[Added] WHEEL BUILDER: This is a new tool available in the Robot Workshop that allows you to create custom wheels.  Try it and see what you think!

[Added] New Robot Workshop Hotkeys:

Added hotkeys for the Electronics - Extras sections, including using wasdqe to use the move/rotate/scale components. When adding components, wasdqe can be used to rotate the component before attaching it to the robot.​

[Added]  33mm Brushless Motor

[Added] New mesh for the AmpFlow A30-400

[Added] New mesh for the AmpFlow F40-300

[Added] New Motenergy motor

[Added] New Heavy Brushless motor

[Added] New Pancake E-Tek motor​

[Added] LET THERE BE LIGHT! Some LED's for the Extras. The lights are baked to reduce strain on the CPU, but despite this they still look really good!

[Added] New Component: Ball Caster. Should eliminate a lot of driving and mobility issues.​​

[Changed] Changed the direction of the NPC and A28 motors 

[Changed] Updated materials for 33mm Brushless​ Motor

[Changed] Tweaked position of AmpFlow A28-150 gearbox so that it sits flush with the motor.​

[Bug Fix] Fixed bug where Blur Cylinders were erroneously added to custom-made wheels attached to drive motors. This was caused because Robot_Reconstruction only checked to see if the previous component had a Control_Motor_ script attached. It now checks recursively up the "previous component" heirarchy to see if any of the parents have the script.

[Bug Fix] Set the hinged rigidbody's angular velocity to zero immediately when all of the attached components break off and only the axle is left.

[Bug Fix] Fixed problem with the Ampflow A28-150 gearbox axle moving out of it's position.

[Bug Fix]Found and fixed the root cause of a series of gyroscopic effect bugs clearly outlined by CodeSilver23 on GTM:



The root cause is that when a rigidbody doesn't contain any colliders, its inertia tensor is set to (1,1,1). This is an ABSOLUTELY MASSIVE set of moments of inertia. When the rigidbody gets up to several thousand RPM, the gyroscopic effects become overwhelming. The fix was to set the MOIs to 0.00001.​

Offline Arcane

  • The Mildly Pretentious
  • Lightweight
  • Posts: 221
  • Rep: 9
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1773 on: February 15, 2020, 10:14:44 AM »
So I see we've reset the sizes of the components added by the previous updates lol

Build Update Issue.png

Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1774 on: February 15, 2020, 10:19:38 AM »
I LOVE the new look of Kat 3!   :laughing

I have no idea what caused that, but would be happy to take a look.  Would you mind sending the .RR2Bot for it?

Offline kix

  • RR2 dev
  • *
  • Posts: 3452
  • Rep: -3
  • H
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1775 on: February 15, 2020, 11:06:15 AM »
The thing is if you are going to fix it, i do not want to resize stuff then

Offline cjbruce

  • Super Heavyweight
  • Posts: 963
  • Rep: 11
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1776 on: February 15, 2020, 11:13:15 AM »
The thing is if you are going to fix it, i do not want to resize stuff then

What sort of things are getting messed up?  Can you see a pattern to it?

Offline Team Lightning

  • Addicted to combat robotics, good at annoying people, and wanting to get to -10000 rep XD.
  • Beetleweight
  • Posts: 168
  • Rep: -4
  • Education is important, robots are importanter.
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1777 on: February 15, 2020, 11:17:46 AM »
Can you rotate the meshes for the ampflow a40-300 and f30-400 180 degrees?

Offline kix

  • RR2 dev
  • *
  • Posts: 3452
  • Rep: -3
  • H
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1778 on: February 15, 2020, 11:22:53 AM »
The thing is if you are going to fix it, i do not want to resize stuff then

What sort of things are getting messed up?  Can you see a pattern to it?

From what ive seen atm, the tri cylinders, hex cylinders  and octo cylinders are all badly sized.

One nitpick is that the new motors dont have the proper attachment points. Now all of them are attachable on the axle which is weird by itself, but new modes of the same motors have their axle points on the side. Ill prolly provide a video

Offline Arcane

  • The Mildly Pretentious
  • Lightweight
  • Posts: 221
  • Rep: 9
    • View Profile
    • Awards
Re: Robot Rumble 2.0 - Robot Combat Simulator - Under Development
« Reply #1779 on: February 15, 2020, 11:25:22 AM »
The thing is if you are going to fix it, i do not want to resize stuff then

What sort of things are getting messed up?  Can you see a pattern to it?

Someone correct me if I’m wrong. It seems to be the 3 non-hollow components added in the previous update; Hexagon/Triangle/The Other One. They all had bizarre base sizes (around 11.0 X/Y/Z) if these have been reset to the same as the rest (1.0) then that would explain why it’s stretched the components out.