Memory pools responsible for models?

ARC_Commander · Post by **ARC_Commander** » Mon Jul 28, 2008 11:56 am

After adding CW to Hoth, I have gained some insight as to the mysterious AT-ST problem. Now, it also affects AT-RTs and droid trifighters. The map seems to strip vehicles of their bones and / or their collision data. (The trifighter can fly through anything, and units can pass through the AT-RT.)

So I'm wondering, could someone provide me with a list of memory pools that have anything to do with models, vehicles and whatnot?

MandeRek · Post by **MandeRek** » Mon Jul 28, 2008 2:51 pm

I've been tweaking my Tridroid walker for months now, since collision didn't work.. It might be my map!

PLEASE someone help!! So, you say a memorypool basicly destroyed a walker's collision? For some reason i'm happy now

ARC_Commander · Post by **ARC_Commander** » Mon Jul 28, 2008 3:05 pm

Yes! That's exactly the problem. Then again, it might just be eliminating a bone from memory or something, taking away the associated collisions with it. Because the AT-ST's legs keep their collisions, but the head loses them.

For me, it occurs on Hoth CW and GCW, and Endor GCW. I don't know what's causing it - I'm just wondering if increasing any certain memory pool would remedy the problem.

I'll try and make my own hex-edit of Hoth, as it might be something ]v[ did that's causing it.

MandeRek · Post by **MandeRek** » Mon Jul 28, 2008 3:09 pm

Can't you post your memory pools? Then I'll do it too, and we see which one we have in common, which isn't, or higher, on one of your modes where they ARE working..

Post by **Teancum** » Mon Jul 28, 2008 3:12 pm

I've never seen/heard of any memory pools affecting vehicle models/animations one way or another. I would check your memory usage to see if you're going beyond the game's limits. In the debug exe, hit ~ and type mem and enter. The usage will show on screen and in your BFront2.log.

ARC_Commander · Post by **ARC_Commander** » Mon Jul 28, 2008 3:28 pm

Here's what I got on Hoth GCW:

Code: Select all

Heaps
  Application: 15684032 free of 15990768
  Gamespy: 2097136 free of 2097136
  Debug: 3145680 free of 6291456
  Lua: 15410396 free of 16777216
  Artist: 8192 free of 8192
  Runtime: 15124816 free of 32235168
    Animation Memory used 811620 bytes
  (null): 0 free of 0
  ProfileView: 501920 free of 1572848
  HUDEditHeap: 957696 free of 1048560
Sound Memory 8224896 free of 33554432
Pool Memory -1024388 free of 4194304
Graphics Memory (8746746 bytes)
   Texture: 0 bytes
   Static Vertex Buffer: 7365604 bytes
   Dynamic Vertex Buffer: 0 bytes
   Index Buffer Memory: 1381142 bytes

CW Hoth got similar results. One thing I'm noticing is the negative pool memory.

Also, what does the "TreeGridStack" pool control? BFront2.log says to increase it.

Post by **Teancum** » Mon Jul 28, 2008 3:33 pm

It's been awhile since I've had to debug some stuff, but your animation memory may be full according to the log. Do you get any animation errors at all in the log?

ARC_Commander · Post by **ARC_Commander** » Tue Jul 29, 2008 10:39 pm

Nope, no animation errors. However, I did notice:

Code: Select all

Message Severity: 2
.\Source\EntityWalker.cpp(4273)
Walker 'rep_walk_atte' skeleton does not contain LegBoneTopLeft 'bone_fl_shoulder'

Message Severity: 2
.\Source\EntityWalker.cpp(4286)
Walker 'rep_walk_atte' skeleton does not contain LegBoneTopRight 'bone_fr_shoulder'

Message Severity: 2
.\Source\EntityWalker.cpp(4210)
Walker 'rep_walk_atte' skeleton does not contain FootBoneLeft 'bone_l_toe'!

Message Severity: 2
.\Source\EntityWalker.cpp(4219)
Walker 'rep_walk_atte' skeleton does not contain FootBoneLeft 'bone_r_toe'!

Message Severity: 2
.\Source\EntityWalker.cpp(4273)
Walker 'rep_walk_atte' skeleton does not contain LegBoneTopLeft 'bone_bl_shoulder'

Message Severity: 2
.\Source\EntityWalker.cpp(4286)
Walker 'rep_walk_atte' skeleton does not contain LegBoneTopRight 'bone_br_shoulder'

Message Severity: 2
.\Source\EntityWalker.cpp(4210)
Walker 'rep_walk_atte' skeleton does not contain FootBoneLeft 'bone_l_toe'!

Message Severity: 2
.\Source\EntityWalker.cpp(4219)
Walker 'rep_walk_atte' skeleton does not contain FootBoneLeft 'bone_r_toe'!

Some stuff appears to be getting forced out of memory. Strangely enough, the two ATTEs are fine.

Edit
Hate to double post, but this is quite important.

I've solved the AT-ST bug, the AT-R bug, and the missing collision bug!

Okay, a long time ago, when I had just started work on BFX, I took a look at the three different AT-ST ODFs - imp_walk_atst, imp_walk_atst_jungle, and imp_walk_atst_snow. What struck me as unusual was this:

Although the three walkers had identical properties except for their meshes, the jungle and snow AT-STs were not class children (eg. with the ClassParent line) of the normal AT-ST. So I thought, hey, why not streamline the whole process, so I made the two AT-STs into class children.

Oh my [Insert Deity Here], it makes so much freakin' sense now! The reason the bug didn't occur on Felucia was because it used the normal AT-ST, which wasn't a class child. The reason the AT-RT and starfighters were affected is because I made child ODFs with shorter names (eg. rep_walk_atrt, cis_fly_dsf) so hex-editing the map would be easier.

So what have I discovered, that Pandemic seems to have discovered several years previous? Having walkers as class children doesn't work.

Does this help, Manderek?

Memory pools responsible for models?

Memory pools responsible for models?

Re: Memory pools responsible for models?

Re: Memory pools responsible for models?

Re: Memory pools responsible for models?

Re: Memory pools responsible for models?

Re: Memory pools responsible for models?

Re: Memory pools responsible for models?

Re: Memory pools responsible for models?