Author Topic: SHW class limit (Read 6934 times)

Venko · « **Reply #20 on:** May 20, 2008, 10:58:30 PM »

Finally a smart one answer about this topic from kill343gs...
And belive me : NO ONE WILL MAKE BOT ABOVE ~8000 or so cos you cant imagine what loading time is this,trust me,50% of crash game or 50% i must wait ~30 min for loading the infinite to edit him on overclocked intel quad core on 3.0 ghz so....
But i think for packs like starcore is important loading time not weight...

70 CUDA · « **Reply #21 on:** May 21, 2008, 12:00:58 AM »

thats funny i have bot that tips the scales at 12.6k with no fridges. so they are done venko. but i am sry i wasted the space on this topic as ideas are hard to get across in a froum. dont missunderstand me i just wanted to cover a space i thought was open but i was worng. thanks for the feedback guys. btw i WILL host a SHW tourny if anyone is in to it.

Naryar · « **Reply #22 on:** May 21, 2008, 03:11:50 AM »

Quote from: venko;8419

Finally a smart one answer about this topic from kill343gs...
And belive me : NO ONE WILL MAKE BOT ABOVE ~8000 or so cos you cant imagine what loading time is this,trust me,50% of crash game or 50% i must wait ~30 min for loading the infinite to edit him on overclocked intel quad core on 3.0 ghz so....
But i think for packs like starcore is important loading time not weight...

Wrong. Go and see Hammer of Justice by Firebeetle. He weights 600 tons. And I've seen him on Youtube... he is perhaps two times the size of the Combat Arena.

___________

For your tourny CUDA, Ill enter. But what is the weight limit?

Madiaba · « **Reply #23 on:** May 21, 2008, 07:14:16 AM »

Venko, It's not mainly the SIZE of the bot/geomobject that really increases lag, it's the total number of VERTICES(angles) that the computer has to calculate. Thus a square arena floor is less calculating than a round wheel on a bot; 8 vertices compared to 30+ respectively. These vertices can be demo-ed by the number of angle points you select to form the shape of your bot in the lab. The comp can handle these no problem, but as you add more geomobjects (and especially rounder ones) the required caculating adds up quickly. So it's not the weight of a SHW that causes lag, but the number of components used to make him a SHW that causes the lag.

Naryar · « **Reply #24 on:** May 21, 2008, 07:37:54 AM »

So basically, the more complex polyhedra you have in collision meshes , the more it lags?

Huh, I just Wiki-ed the thing and I now know the meaning of "Vertex"...

Madiaba · « **Reply #25 on:** May 21, 2008, 07:47:38 AM »

Yea, Naryar. And not just Collision meshes, but also Display meshes.
Thus the 'Battle Box' arena is much "laggier" than the giant 'Empty' arena which is perhaps 10x larger.
When I build arenas (and components), I strive to make objects as simple as possible while maintaining aesthetics. This leaves more 'computer calculating' for the bots and their functionings.
ACAMS did the same with the USS Enterprise.gmf he made for me.

Venko · « **Reply #26 on:** May 21, 2008, 07:51:43 AM »

did you read this in my previous post?

"I found also that not weight make bots loading a lot but puting many parts on one part... For example the infinite was loading for ~2 min almos ready and when i just add few spinners now is 30 min so...
Complicated weapons make lag not many parts or weight...
Why dont you try to make bots with many parts and big weight,but without many parts attached to each other and will see what happen then...
For example bots with complicated weapons like bots that are with waepons close to "no more than 7 parts att. rule" are too lagy even that they are HW!"

How can you explain it...

Madiaba · « **Reply #27 on:** May 21, 2008, 08:27:37 AM »

Quote from: venko;8430

did you read this in my previous post?
"I found also that not weight make bots loading a lot but puting many parts on one part..

Your point here is good and true, Venko,... but only in part.

Quote from: venko;8430

Complicated weapons make lag not many parts or weight.... Why dont you try to make bots with many parts and big weight,but without many parts attached to each other and will see what happen then...

This is the part that isn't true.
Explanation:
(1) In the bot file, the comp loads the same amount of info whether a component is attached to the base of the bot or another component. So 7 consecutive components attached in a line, bears the same amount of info to crunch as 7 components attached to the bot's base. Total number of vertices is the same too.
(2) But there IS a difference in computer load between 7 consecutive components and 7 base attached components IF the components are 'dynamic' (motors, pistons,...). But this has little to do with the total number of vertices. When attaching a dynamic component to another dynamic component (motor/piston/servo to another motor/piston/servo) both Havoc and the comp work allot harder calculating each of the physical properties, dynamics, and effects of the series; and the farther away from the first, the more complicated to calculate (exponentially, IMO). Thus you can easily have 4 pistons attached to the base of your bot, but not 4 pistons atached to one another without functioning problems. Same with spinmotors, servos, etc....
BTW: This is what enables many assembly glitches.

Flying_Chao · « **Reply #28 on:** May 21, 2008, 02:07:01 PM »

Quote from: 70 CUDA;8421

thats funny i have bot that tips the scales at 12.6k with no fridges. so they are done venko. but i am sry i wasted the space on this topic as ideas are hard to get across in a froum. dont missunderstand me i just wanted to cover a space i thought was open but i was worng. thanks for the feedback guys. btw i WILL host a SHW tourny if anyone is in to it.

What would the limit be, if there was one?

70 CUDA · « **Reply #29 on:** May 21, 2008, 02:12:36 PM »

3.25K guys if you want to be in it then it will be after my tourny.

Naryar · « **Reply #30 on:** May 21, 2008, 03:17:16 PM »

ZOMG! i'd have to enter a 3,25K VS! l0lz!

Can I suggest a few rules?

-No more than two different weapon types (example, hybrid Rammer/HS is okay, but dont add a hammer on it), this for sparing CUDA's sanity.

-No multibots, for the same reason(that'll make 4 SHWs, three against one, in the arena...) and for decreasing lag

-Bots must show controlled movement at all times, but non-wheeled things are allowed if they can drive and steer at least in a bad way (so controlled bumble bees are in, happy Venko?)

-Fights will last 5 minutes, 'cause theres no such thing as a 1 hit kill in SHWs.

70 CUDA · « **Reply #31 on:** May 21, 2008, 03:22:30 PM »

ok all that seems good. have fun making them guys

Flying_Chao · « **Reply #32 on:** May 21, 2008, 03:22:37 PM »

I don't think 5 minutes is long enough for a SHW battle....

70 CUDA · « **Reply #33 on:** May 21, 2008, 03:28:32 PM »

ill take a look in the the time of each match

Naryar · « **Reply #34 on:** May 21, 2008, 03:38:35 PM »

Quote from: Flying_Chao;8443

I don't think 5 minutes is long enough for a SHW battle....

Good point. I was thinking that too, but put 5 minutes...

Well O.K. it will be 10 minutes then, but beware the batteries...

So Ill add:

-If a havok incurs and gives a CLEAR help to one of the bots, then restart the match (as BBEANS)

-Match time is 10 minutes.

What about the arena? Itll be fun to make that on the Octagon, but not very practical. The Combat Arena looks like the solution.

Flying_Chao · « **Reply #35 on:** May 21, 2008, 03:45:55 PM »

If you turn off the hazards, the octagon might work, but the combat arena is the usual fit.

kill343gs · « **Reply #36 on:** May 21, 2008, 06:22:17 PM »

Or you could get a slightly bigger arena like the huge arena or something, but theres the problem of the high fall spawn point.

Flying_Chao · « **Reply #37 on:** May 21, 2008, 08:43:26 PM »

Wouldn't it also be hard to get the camera close to where the bots are actually fighting?

Trovaner · « **Reply #38 on:** May 21, 2008, 09:01:48 PM »

You could use the Infinite Arena (so that it the bots have normal starting points and because it is very low lag) with the F10 camera.

70 CUDA · « **Reply #39 on:** May 22, 2008, 12:18:27 PM »

ok i might be able to fix the high starting point, but so far i like what im seeing. thanks for your input