dim3 Forum

(2008.09.15 08:13 PM)cyst Wrote: [ -> ]I'm guessing that the "gigs of textures" people are the ones who are being stupid and trying to bake lighting for entire maps. Nothing is going to kill performance like trying to create an entire course with no repeating textures -- and that's true in any engine on the market.

ID's Tech 5 engine can do it without destroying performance- in fact the artists can go back and individually alter every single texture and still have solid performance. Brad beat me to it, though.

(2008.09.15 08:32 PM)Bradamante Wrote: [ -> ]

(2008.09.15 08:13 PM)cyst Wrote: [ -> ]and that's true in any engine on the market.

You sure know what you are talking about.

You're comparing dim3 to id tech 5? Seriously?

(2008.09.16 12:05 AM)cyst Wrote: [ -> ]You're comparing dim3 to id tech 5? Seriously?

awwwwright, I won't make that comparison. I think that a slightly fairer comparison for dim3 would be Unity. Unity is superior in a number of ways, but then again they have a real team of people working on it, wheras Dim3 only has Brian. I like Dim3 more though.

Let's remember that unlike most engines, I'm attempting to keep the system requirements as low as possible. There are some amazing engines out there, but there are also some incredibly high system requirements, and nobody has really gotten hands on with tech 5 yet. Engines are almost always specifically good at something, very free are jack of all trades. There's always a pay off But this is off topic ...

I think when I finally finish with the all-mesh version, especially in the free non-professional category -- I'll be pretty close to most of the ones out there.

[>] Brian

---

Next version!

Same Location: http://homepage.mac.com/ggadwa/Misc/MeshOnlyTest.zip

Changes:


IMPORTANT NOTES!

* editor is very slow, I know this, it's not high priority right now but I'll get to it
* obscuring works better now but it's not perfect. The better I make it, the slower it gets. Therefore, I'm probably going to make a map switch where you can do a "show all meshes always", "rough", or "complete" obscure calculations, to make map editing faster.
* if your map crashed before, you will NEED TO SAVE IT IN EDITOR FIRST. Editor now fixes problems that would crash maps in Engine
* Engine only runs current maps now, old maps need to be run through Editor before using them. There's too much new calculation that Editor needs to do
* I want to reduce memory requirements a bit for the next one

MAIN GOAL!

No crashing in Engine. I want to get that part solid.

Should be getting close with the editors, then I'm on to physics and networking fixes.

[>] Brian

Still crashes for me on maps w/ alpha-channel textures. Replacing the transparent textures with opaque ones fixes the problem. Changing them back to different alpha-channel textures causes the problem to come back.

Also, jumping still causes the player to fall through the floor.

Update: Crash Log:

Brian, I keep encountering a problem where the map never fully loads. Sometimes it starts playing audio from the map, but the engine never quite gets past the loading screen. I save the maps in editor and run them in the engine...excep the loading bar gets about 99% done and doesn't finish.

(2008.09.16 12:50 AM)cyst Wrote: [ -> ]Still crashes for me on maps w/ alpha-channel textures. Replacing the transparent textures with opaque ones fixes the problem. Changing them back to different alpha-channel textures causes the problem to come back.

Also, jumping still causes the player to fall through the floor.

Update: Crash Log:

Code:

Process:         dim3 Engine [7219]
Path:            /Users/[redacted]/Downloads/MeshOnlyTest/dim3 Engine.app/Contents/MacOS/dim3 Engine
Identifier:      com.klinksoftware.dim3.engine
Version:         dim3 Engine version 2.5 (2.5)
Code Type:       X86 (Native)
Parent Process:  launchd [310]

Date/Time:       2008-09-15 21:55:04.512 -0700
OS Version:      Mac OS X 10.5.4 (9E17)
Report Version:  6

Exception Type:  EXC_BAD_ACCESS (SIGSEGV)
Exception Codes: KERN_INVALID_ADDRESS at 0x00000000262b9e3c
Crashed Thread:  Unknown

Thread 0:
0   libSystem.B.dylib                 0x9667768e __semwait_signal + 10
1   libSystem.B.dylib                 0x966a236d pthread_cond_wait$UNIX2003 + 73
2   libGLProgrammability.dylib        0x91ef7432 glvmDoWork + 162
3   libSystem.B.dylib                 0x966a16f5 _pthread_start + 321
4   libSystem.B.dylib                 0x966a15b2 thread_start + 34

Thread 1:
0   libSystem.B.dylib                 0x966704a6 mach_msg_trap + 10
1   libSystem.B.dylib                 0x96677c9c mach_msg + 72
2   com.apple.CoreFoundation          0x94ba30ce CFRunLoopRunSpecific + 1790
3   com.apple.CoreFoundation          0x94ba3cf8 CFRunLoopRunInMode + 88
4   com.apple.audio.CoreAudio         0x94833464 HALRunLoop::OwnThread(void*) + 160
5   com.apple.audio.CoreAudio         0x94833300 CAPThread::Entry(CAPThread*) + 96
6   libSystem.B.dylib                 0x966a16f5 _pthread_start + 321
7   libSystem.B.dylib                 0x966a15b2 thread_start + 34

Thread 2:
0   libSystem.B.dylib                 0x96670506 semaphore_timedwait_signal_trap + 10
1   libSystem.B.dylib                 0x966a284f _pthread_cond_wait + 1244
2   libSystem.B.dylib                 0x966a40d3 pthread_cond_timedwait_relative_np + 47
3   com.apple.audio.CoreAudio         0x94842a47 CAGuard::WaitFor(unsigned long long) + 213
4   com.apple.audio.CoreAudio         0x94844602 CAGuard::WaitUntil(unsigned long long) + 70
5   com.apple.audio.CoreAudio         0x94842da7 HP_IOThread::WorkLoop() + 759
6   com.apple.audio.CoreAudio         0x94842aab HP_IOThread::ThreadEntry(HP_IOThread*) + 17
7   com.apple.audio.CoreAudio         0x94833300 CAPThread::Entry(CAPThread*) + 96
8   libSystem.B.dylib                 0x966a16f5 _pthread_start + 321
9   libSystem.B.dylib                 0x966a15b2 thread_start + 34

Unknown thread crashed with X86 Thread State (32-bit):
  eax: 0x00000000  ebx: 0x00010a04  ecx: 0x1b4a9020  edx: 0x200d45d4
  edi: 0x0000162b  esi: 0x1b542280  ebp: 0x262a9430  esp: 0xbfffe7e8
   ss: 0x0000001f  efl: 0x00010246  eip: 0x1a52980b   cs: 0x00000017
   ds: 0x0000001f   es: 0x0000001f   fs: 0x00000000   gs: 0x00000037
  cr2: 0x262b9e3c

Cyst -- Can you put up your project (with the crashing map), and also tell me where in the map to jump up to get the problem?

Kieranator -- which of the maps that I have does this?

[>] Brian

(2008.09.16 09:16 AM)ggadwa Wrote: [ -> ]Kieranator -- which of the maps that I have does this?

[>] Brian

Uh, none of them anymore. O_o

But there are a LOT of obscuring problems on Kastle and performance on Kastle is about 0.1 frames per second on an Nvidia 8600 GT. It's about 0.2 frams per second on Sewers, and about 2 frames per second on Arena. I think that something needs to happen to get performance back to about where it was with the old engine.

(2008.09.16 09:16 AM)ggadwa Wrote: [ -> ]Cyst -- Can you put up your project (with the crashing map), and also tell me where in the map to jump up to get the problem?

Here is a data folder that illustrates the problems I'm having.

The Arena1 map has only one alpha-blended texture (texture #9). When any of the meshes that use it become visible, the engine crashes. Replacing it with an opaque texture solves the crashing problem; replacing it with a different transparent texture does not. Setting the opacity of any polygon to anything less than 1 causes the same problem. Also, the room where the player starts illustrates the jumping problem, as does the room with the stairs.

The Atrium map is one that was imported from the old engine, and has, I think, more complex problems (I've sent you this map before ). It crashes on launch, but replacing the transparent textures doesn't solve the problem (this map has not worked right in any version since b9; maybe I should just scrap it?)

(2008.09.16 09:34 AM)Kieranator Wrote: [ -> ]

(2008.09.16 09:16 AM)ggadwa Wrote: [ -> ]Kieranator -- which of the maps that I have does this?

[>] Brian

Uh, none of them anymore. O_o

But there are a LOT of obscuring problems on Kastle.

-K

OK, so let me get this straight -- no more crashes (for you) -- but just obscuring problems. I know I need work there, so I'll check out your map.

[>] Brian