CoreDumping/Massive Error messages while tracing.

CoreDumping/Massive Error messages while tracing.

Post by Garvin Hicki » Fri, 11 Jul 1997 04:00:00



Hi!

I have a serious problem with a current POVRay image of mine. It's one
for the IRTC and uses the following "features":

- many bezier-patches
- two halos
- focal blur
- transmitting textures
- many, many, CSG-Differences/Intersections: About 150 Objects
stamping through one single, flat cube.
- some lathe-objects

I have tried to render it in POVRay3.01.dos.watcom, 3.01.dos.dos-4g,
3.02.dos.watcom, 3.02.linux and 3.01e.windows. In all versions, the
image crashed after about 50-100 lines (at 800x600). Under DOS it
screwed up with the following message:
--------------------------------snip-----------------------------------------------
CauseWay DOS Extender v3.14 Copyright 1992-96 Michael Devore.
All rights reserved.

Exception: 0E, Error code: 0004

EAX=005083E8 EBX=00000000 ECX=004D6308 EDX=40A52CA5 ESI=00000000
EDI=00000000 EBP=00508414 ESP=005083C8 EIP=0049C2E8 EFL=00013201

CS=01A7-FFBCA000 DS=01AF-FFBCA000 ES=01AF-FFBCA000
FS=0000-xxxxxxxx GS=01B7-xxxxxxxx SS=01AF-FFBCA000

CR0=00000000 CR2=00000000 CR3=00000000 TR=0000

Info flags=00008040

$
CS:EIP > 8B 52 14 E8 04 75 FC FF 85 C0 74 6C 8B 44 24 44
         6B 40 08 58 8B 54 24 44 8B 52 04 8B 0C 1C 89 0C
         02 8B 4C 1C 04 89 4C 02 04 8B 44 24 44 6B 50 08
         58 8B 40 04 01 D0 8B 54 24 3C 89 50 38 8B 44 24
         44 6B 40 08 58 8B 54 24 44 8B 52 04 B9 18 00 00
         00 01 D0 8D 74 24 20 8D 78 08 57 89 C8 C1 E9 02
         F2 A5 8A C8 80 E1 03 F2 A4 5F 8B 44 24 44 BE 01
         00 00 00 E8 98 78 FC FF 8B 54 24 38 83 C3 08 39
         D3 0F 8C 1F FF FF FF 89 F0 89 EC 5D 5F 5E 59 C3
         51 56 57 55 89 E5 81 EC D8 00 00 00 83 E4 F8 89
         C6 89 D1 89 DF 8B 15 4C A1 4E 00 42 89 15 4C A1
         4E 00 75 06 FF 05 48 A1 4E 00 8D 44 24 60 89 F2
         8B 59 34 E8 D0 1E FC FF 8D 44 24 48 8D 56 18 8B
         59 34 E8 91 1F FC FF DD 44 24 50 D8 C8 DD 44 24
         48 D8 C8 DE C1 DD 44 24 58 D8 C8 DE C1 D9 FA D9
         E8 DD 44 24 48 D9 CA DD 9C 24 D0 00 00 00 DC B4

SS:ESP > 7E C8 6E 44 BA 2D A5 40 5E 01 42 52 B6 2D A5 40
         1D 02 36 CE 5E 2D A5 40 9F 6E 36 B3 2C 2D A5 40
         00 BB D3 36 6A 82 00 C0 00 B8 9E FF FA 1A 01 40
         C0 FF C1 5F D8 A4 DF 3F 40 00 00 00 A5 2C A5 40
         48 89 50 00 30 1F 7B 00 11 11 1F F5 38 92 4D 00
         00 00 00 00 70 AB 13 01 08 63 4D 00 AA 70 44 00
         48 89 50 00 AB 01 A2 FB 00 00 00 00 30 1F 7B 00
         08 63 4D 00 00 00 00 00 80 A6 13 01 08 63 4D 00
         AA 70 44 00 48 89 50 00 25 CB 39 C8 00 00 00 00
         30 1F 7B 00 08 63 4D 00 00 00 00 00 48 94 13 01
         08 63 4D 00 AA 70 44 00 48 89 50 00 CD 42 0C 9E
         00 00 00 00 30 1F 7B 00 30 1F 7B 00 48 89 50 00
         00 00 00 00 08 63 4D 00 4B 73 44 00 48 89 50 00
         00 00 00 00 D0 14 7B 00 F0 93 13 01 B8 4B 51 00
         18 21 09 01 D4 84 50 00 48 89 50 00 30 63 4D 00
         D0 14 7B 00 C9 36 46 00 30 63 4D 00 E8 84 50 00

         3B CF D2 40 00 00 10 59 3C CF D2 40 E4 DF DB 30
         57 CF D2 40 09 ED 65 B2 83 CF D2 40 CB 29 B8 6E
         D1 CF D2 40 94 02 01 F0 D4 CF D2 40 D8 82 56 E1
         E9 37 6A 40 18 EB C3 43 61 48 D2 C0 AF 8D D0 C0
         3D 7B E5 40 B2 58 32 40 70 BC 55 00 0E CE 8B 12
         70 1D B6 41 A4 EC 8E CD 01 43 02 C0 DB A4 C2 EB
         BE E6 EA 3F 00 00 00 00 00 00 02 40 7C 14 AE 47
         E1 7A A4 3F 0B D7 A3 70 3D 0A FB 3F 9A 99 99 99
         99 99 C9 3F 1E 85 EB 51 B8 1E 07 40 9A 99 99 99
         99 99 C9 BF D0 A7 00 65 84 6C 1C 40 AF 01 00 00
         00 00 00 00 E8 83 50 00 35 06 00 00 27 01 00 00
         01 32 01 00 7E C8 6E 44 BA 2D A5 40 5E 01 42 52
         B6 2D A5 40 1D 02 36 CE 5E 2D A5 40 9F 6E 36 B3
         2C 2D A5 40 00 BB D3 36 6A 82 00 C0 00 B8 9E FF
         FA 1A 01 40 C0 FF C1 5F D8 A4 DF 3F 40 00 00 00
         A5 2C A5 40 48 89 50 00 30 1F 7B 00 11 11 1F F5
SS:EBP > 38 92 4D 00 00 00 00 00 70 AB 13 01 08 63 4D 00
         AA 70 44 00 48 89 50 00 AB 01 A2 FB 00 00 00 00
         30 1F 7B 00 08 63 4D 00 00 00 00 00 80 A6 13 01
         08 63 4D 00 AA 70 44 00 48 89 50 00 25 CB 39 C8
         00 00 00 00 30 1F 7B 00 08 63 4D 00 00 00 00 00
         48 94 13 01 08 63 4D 00 AA 70 44 00 48 89 50 00
         CD 42 0C 9E 00 00 00 00 30 1F 7B 00 30 1F 7B 00
         48 89 50 00 00 00 00 00 08 63 4D 00 4B 73 44 00
         48 89 50 00 00 00 00 00 D0 14 7B 00 F0 93 13 01
         B8 4B 51 00 18 21 09 01 D4 84 50 00 48 89 50 00
         30 63 4D 00 D0 14 7B 00 C9 36 46 00 30 63 4D 00
         E8 84 50 00 D0 14 7B 00 00 00 00 00 00 00 F0 7F
         A8 85 50 00 08 00 00 00 02 00 00 00 40 85 50 00
         01 99 43 00 DF 0E 83 C0 AD 67 DB 3F 31 70 0C 82
         EC DD 4E 40 FC 72 66 BB C2 D4 65 40 80 D2 B1 FE
         08 85 50 00 31 70 0C 82 EC DD 4E 40 FC 72 66 BB

   Application resource tracking details  
===========================================

Selectors
=========

 sel   base     limit  type D mem count
----------------------------------------
018F 00431000 00004000 DATA 32 Y xxxx
019F 00431AEC 00000264 DATA 32 Y xxxx
01A7 00000000 FFFFFFFF CODE 32 N 0001
01AF 00000000 FFFFFFFF DATA 32 N 0002
01B7 00431D94 00000043 DATA 32 Y xxxx

Total selectors: 0005

Linear memory blocks
====================

 handle    base    length
--------------------------
00436000 00436000 000D6000
00431AEC 00431AEC 00000268
00431D5C 00431D5C 00000018
00431D94 00431D94 00000044

Total Linear memory: 000D62C4 (000DA000) in 00000004 blocks

Linear memory locked
====================

  base    length
-----------------
DE5C004B 0000000B

DOS memory blocks
=================

 sel   base    length
----------------------

Protected mode interrupt vectors
================================

No sel   offset
----------------

Protected mode exception vectors
================================

No sel   offset
----------------

Real mode interrupt vectors
===========================

No seg offset
-------------

Call-Backs
==========

  real        target
-----------------------
--------------------------------snip-----------------------------------------------

Under Linux and Unix it just core-dumped with a segmentation fault.
This problem is machine-independent and I am somehow annoyed, because
I really wanted to submit the image to the IRTC.

Does anyone know, where the bug could be?

Many thanks in advance!!!

---

Garvin            -=[ http://home.pages.de/~garvin/ ]=-

 
 
 

CoreDumping/Massive Error messages while tracing.

Post by Chris Colefa » Sat, 12 Jul 1997 04:00:00



> Hi!

> I have a serious problem with a current POVRay image of mine. It's one
> for the IRTC and uses the following "features":

> - many bezier-patches
> - two halos
> - focal blur
> - transmitting textures
> - many, many, CSG-Differences/Intersections: About 150 Objects
> stamping through one single, flat cube.
> - some lathe-objects

> I have tried to render it in POVRay3.01.dos.watcom, 3.01.dos.dos-4g,
> 3.02.dos.watcom, 3.02.linux and 3.01e.windows. In all versions, the
> image crashed after about 50-100 lines (at 800x600). Under DOS it
> screwed up with the following message:

[snip]

> Under Linux and Unix it just core-dumped with a segmentation fault.
> This problem is machine-independent and I am somehow annoyed, because
> I really wanted to submit the image to the IRTC.

> Does anyone know, where the bug could be?

In my experience, these kinds of crashes seem to be caused by
mathematical problems within the raytracer itself (eg. division by 0),
or memory allocation errors.  I have come across a number of animations
that cause a similar crash, but restarting the render from the faulty
frame generally solves the problem.  The same is often the case with
still scenes, so my first advice would be to use the +C option and see
if this allows you to finish the rendering.

Otherwise, I have found that having a large number of objects that all
occupy the same point of space within the POVRay universe also causes a
crash (sort of like the coincident surfaces from hell!).  In this case,
simply separating the objects by a small amount should usually fix the
problem.

If all else fails, you will probably have to start removing sections of
the scene (eg. remove the focal blur, or the lathes), until you can find
the cause of the error.

 
 
 

CoreDumping/Massive Error messages while tracing.

Post by Christopher Shami » Sun, 13 Jul 1997 04:00:00



> that cause a similar crash, but restarting the render from the faulty
> frame generally solves the problem.  The same is often the case with
> still scenes, so my first advice would be to use the +C option and see
> if this allows you to finish the rendering.

Could this be a stack size thing?  or a max_intersections limit?  Anyone
know some good numbers?  Memory I got, (64MB worth), so what would be a good
value so that I'd never *out?

Quote:> Otherwise, I have found that having a large number of objects that all
> occupy the same point of space within the POVRay universe also causes a
> crash (sort of like the coincident surfaces from hell!).  In this case,

Hey!  This worked for me!  It picked *RIGHT* back up.  Bummer, wonder where
it puked.  If any de*s want a scene that will cause this mystery
seg-fault.  Lemme know, cause I've got one, that only has a few objects, (Of
course, they're *GLASS* so it renders like a dog...

--
C. Shamis                                       / I used to have something

                                                / to use this space to tell
people, please note the "REMOVE_THIS" tag in my email address.  Damn spammers!

 
 
 

CoreDumping/Massive Error messages while tracing.

Post by Garvin Hicki » Mon, 14 Jul 1997 04:00:00


Hi Chris!

Thanks for your Hints - I found out, that continuing the scenes works for  
only 5 or 6 lines, and then it crashes again. So I will have to run it a  
lot, but I think somehow I can finish the image.

Otherwise I will try your suggestions about moving the objects a bit  
around, or maybe even scale them larger.

CC> If all else fails, you will probably have to start removing sections of
CC> the scene (eg. remove the focal blur, or the lathes), until you can
CC> find the cause of the error.

That would be the hardest part of all... :-|

---

Garvin                                -[= http://home.pages.de/~garvin/ =]-
--
|FidoNet : Garvin Hicking 2:2453/690.16

|
|From IceMan's BBS, Bad Honnef, Germany [+49-2224-931979]