SGI: Computer Graphics

Blender Irix Development - Page 5

Jarndyce911 wrote:
Actually I had to recompile once I got the new version of OpenEXR (which is used in Blender), I just forgot to upload the new version of Yafray. Sorry about that (I have been busy).

This should be it:

http://www.nekochan.net/wiki/nekoware/beta/n ... 9a.tardist

Please let me know if it still does not work!

Thanks,


Cool, thanks for that!

The library problem is solved but I'm getting a crash when trying a test render of the 'big.xml' file dexter1 posted a while back:

Code:
neko@Mai ~ % yafray big.xml
Starting YafRay ...
Loading grammar ...
Starting parser ...
Parsing OK

Loading plugins from '/usr/nekoware/lib/yafray'...
Registered arealight
Registered basicblocks
Registered basicshaders
Registered blendershaders
Registered globalphotonlight
Registered HDRI background
Registered hemilight
Registered mix block
Registered pathlight
Registered photonlight
Registered pointlight
Registered Shader Background
Registered softlight
Registered spherelight
Registered spotlight
Registered sss
Registered sunlight
Registered sunsky
found 18 plugins!
[Loader]: Added shader MAMaterial
[Loader]: Added shader MAcristal
[Loader]: Added shader MArojo
starting build of kd-tree

=== kd-tree stats (0.22s) ===
primitives in tree: 3328
interior nodes: 9111 / leaf nodes: 9112 (empty: 4397 = 48.255%)
[Loader]: Added object OBCube.001
starting build of kd-tree

=== kd-tree stats (0s) ===
primitives in tree: 12
interior nodes: 22 / leaf nodes: 23 (empty: 9 = 39.1304%)
[Loader]: Added object OBCube.002
starting build of kd-tree

=== kd-tree stats (0s) ===
primitives in tree: 12
interior nodes: 25 / leaf nodes: 26 (empty: 10 = 38.4615%)
[Loader]: Added object OBCube.003
starting build of kd-tree

=== kd-tree stats (0s) ===
primitives in tree: 2
interior nodes: 1 / leaf nodes: 2 (empty: 1 = 50%)
[Loader]: Added object OBPlane
starting build of kd-tree

=== kd-tree stats (0s) ===
primitives in tree: 12
interior nodes: 1 / leaf nodes: 2 (empty: 1 = 50%)
[Loader]: Added object OBCube
[Loader]: Added photonlight light LAMP1
[Loader]: Added sunlight light LAMP2
[Loader]: Added pathlight light path_LT
[Loader]: Added constant background world_background
[Loader]: Added camera MAINCAM
Using a world resolution of 0.00142857 per unit
Rendering with 5 raydepth
2 anti-alias passes and 4 minimum samples per pass, 8 samples total.
Building bounding tree ... Object count= 5
OK
Light setup ...
Setting up lights ...
Shooting photons ... OK
Emitted 250000 Stored 170253 search 100
Pre-Gathering (49915) ... Assertion failed in file "../../libC/lang_support/unwind.cxx", line 579
zsh: IOT instruction (core dumped)  yafray big.xml


The binary is stripped so I wasn't able to get a dbx dump.

_________________
Twitter: @neko_no_ko
IRIX Release 4.0.5 IP12 Version 06151813 System V
Copyright 1987-1992 Silicon Graphics, Inc.
All Rights Reserved.
When using Blender to render a scene with the newly compiled yafray it does seem to work.
The only thing I noticed was that when the scene uses an HDMI environment map it doesn't show up in the render.
In the Blender scene however the map is clearly present and shows in the world material preview pane.

This could also be due to new setting variables in blender of which I'm not aware of .... yet...

_________________
I can feel it, my mind is going ....
Blender 2.44 has been released! Here is the link to the Irix version!

http://download.blender.org/release/Ble ... ps.tar.bz2

_________________
Charles
Sweet! I'm gonna try it first thing in the morning - if I turn on the SGI now it'll wake up the misses.

_________________
Maverick 3: Athlon X2 7750, 2gb, Windows Vista
Frank Dux: SGI Octane2 R12k 400mhz 1.5gb

"Chief, look! I learned to make fire! Who knows what we could do with this... We should learn to control it!"

"Ridiculous. How can you justify wasting time and effort on this so-called 'fire' when our children are freezing to death at night?"
Well... i own a small Indy, but this thing load in 2 secs and less ! shure ! Even on an Indy i can manage smaller objects and meshes with np and quite fast ! ... who optimised this thing done it like a pro ! regards ! off course rendering part on an Indy is not a song, but shure ppl with Octanes and on will get max. from this baby :) ...

Today i imported some dxf`s and played a little ... maybe will post some examples here if i will have time ... i never used blender before ... but this one rocks !!!

The problem with importing files, and missing file names is fixed, the libiconv problem is fixed too, but i get an warning window which say that i must adjust the path to $phython home, but i not know where to set it up ...

_________________
Image
-----------------
Image Image Image
nice nice nice

but why be limited to 8 thread..... 2.43 have 1024 threads.....

SSS rocks

regards

Laurent

_________________
SGI or die !!!
:O2: :Octane2: :Octane: :Indigo2IMP: :Indigo2IMP: :Indigo: :Indigo: :Indy: :PI: :Crimson: :PWRSeries: :Onyx: :O2000R:
HP proliant DL 585 Quad Opteron dual core 2.5Ghz 16Gb
fzalfa wrote:
but why be limited to 8 thread..... 2.43 have 1024 threads.....


Actually 2.43 was also limited to 8 threads, I just upped the limit of what you could tell it to do (it could never actually do it though - too many core Blender modifications would be required).

Glad it is working so far, Irix is on the main download page now!

_________________
Charles
Jarndyce911 wrote:
Glad it is working so far, Irix is on the main download page now!


Cool! While I was there I pre-ordered the new Blender guide too :)

_________________
Twitter: @neko_no_ko
IRIX Release 4.0.5 IP12 Version 06151813 System V
Copyright 1987-1992 Silicon Graphics, Inc.
All Rights Reserved.
There's something wrong with Yafray ideed....
When I try to render the example cornell.blend I get the following error:
(This is not the complete output just the last few lines... )
Code:
Using a world resolution of 0.000892857 per unit
Rendering with 5 raydepth
4 anti-alias passes and 4 minimum samples per pass, 16 samples total.
Building bounding tree ... Object count= 4
OK
Light setup ...
Setting up lights ...
Yafray crashed
Could not execute yafray. Is it in path?


Is this because of the new compile of Yafray?

_________________
I can feel it, my mind is going ....
I found out today that when you disable xml output in the Yafray renderoption tab in blender, it does work!
That is, it will render the cornell.blend testfile and makes it look very good ( as usual) :)

I also discovered that when rendering the same scene on a Macbook ( intell inside... ) it looks very different from the one I've rendered using an Octane2.
In my scene there's a moving object made of glass, and on the SGI it looks really realistic, but on
my Macbook it looks like the glass has a foggy texture.

Can anyone do a test with a Yafray render using blender on SGI and another platform and confirm this?

So although my SGI takes a whole lot longer to render then my OS-X machine I still prefer it cause the final result looks so much better !

Thank you Charles for keeping Blender alive for us geeks :wink:

_________________
I can feel it, my mind is going ....
Yeah, the XML button makes it write the data to a file instead of actually rendering immediately.

I tried the new update! It works fantastic, didn't have hte problems I had with 2.43.

Only thing left is I can't change the font for the text tool.

_________________
Maverick 3: Athlon X2 7750, 2gb, Windows Vista
Frank Dux: SGI Octane2 R12k 400mhz 1.5gb

"Chief, look! I learned to make fire! Who knows what we could do with this... We should learn to control it!"

"Ridiculous. How can you justify wasting time and effort on this so-called 'fire' when our children are freezing to death at night?"
With version 2.44 I get more info in the command-line from yafray, why it's not using a background image.
Here's some of the output:
Code:
Rendering with 8 raydepth
4 anti-alias passes and 4 minimum samples per pass, 16 samples total.
[Warning]: Background world_background does not exist
Building bounding tree ... Object count= 15
OK

Somehow it no longer seems to support HDRI images as a background .... !
The same scene worked just fine with version 2.42 ...... :shock:
Can someone confirm this ?

_________________
I can feel it, my mind is going ....
nekonoko wrote:
The library problem is solved but I'm getting a crash when trying a test render of the 'big.xml' file dexter1 posted a while back:

Code:
neko@Mai ~ % yafray big.xml
Starting YafRay ...
Loading grammar ...
Starting parser ...
Parsing OK

Loading plugins from '/usr/nekoware/lib/yafray'...
Registered arealight
Registered basicblocks
Registered basicshaders
Registered blendershaders
Registered globalphotonlight
Registered HDRI background
Registered hemilight
Registered mix block
Registered pathlight
Registered photonlight
Registered pointlight
Registered Shader Background
Registered softlight
Registered spherelight
Registered spotlight
Registered sss
Registered sunlight
Registered sunsky
found 18 plugins!
[Loader]: Added shader MAMaterial
[Loader]: Added shader MAcristal
[Loader]: Added shader MArojo
starting build of kd-tree

=== kd-tree stats (0.22s) ===
primitives in tree: 3328
interior nodes: 9111 / leaf nodes: 9112 (empty: 4397 = 48.255%)
[Loader]: Added object OBCube.001
starting build of kd-tree

=== kd-tree stats (0s) ===
primitives in tree: 12
interior nodes: 22 / leaf nodes: 23 (empty: 9 = 39.1304%)
[Loader]: Added object OBCube.002
starting build of kd-tree

=== kd-tree stats (0s) ===
primitives in tree: 12
interior nodes: 25 / leaf nodes: 26 (empty: 10 = 38.4615%)
[Loader]: Added object OBCube.003
starting build of kd-tree

=== kd-tree stats (0s) ===
primitives in tree: 2
interior nodes: 1 / leaf nodes: 2 (empty: 1 = 50%)
[Loader]: Added object OBPlane
starting build of kd-tree

=== kd-tree stats (0s) ===
primitives in tree: 12
interior nodes: 1 / leaf nodes: 2 (empty: 1 = 50%)
[Loader]: Added object OBCube
[Loader]: Added photonlight light LAMP1
[Loader]: Added sunlight light LAMP2
[Loader]: Added pathlight light path_LT
[Loader]: Added constant background world_background
[Loader]: Added camera MAINCAM
Using a world resolution of 0.00142857 per unit
Rendering with 5 raydepth
2 anti-alias passes and 4 minimum samples per pass, 8 samples total.
Building bounding tree ... Object count= 5
OK
Light setup ...
Setting up lights ...
Shooting photons ... OK
Emitted 250000 Stored 170253 search 100
Pre-Gathering (49915) ... Assertion failed in file "../../libC/lang_support/unwind.cxx", line 579
zsh: IOT instruction (core dumped)  yafray big.xml


The binary is stripped so I wasn't able to get a dbx dump.


I tried this on my Octane2 with Irix 6.5.22 and I didn't have any problems rendering this.
Here's the final render :


Attachment:
big.jpg
big.jpg [ 61.93 KiB | Viewed 526 times ]


And in order to try to keep this software alive for the Irix platform can anyon else give this render a try?

_________________
I can feel it, my mind is going ....
I'd love to, but my gear's all packed up for the moment :)
Maybe in a few months...

_________________
-= I reject reality, and substitute my own =-

1 Indigo R3k-33 32MB XS24-Z;
1 Indy R5k-180 256MB XZ;
1 Indy R4k-175 64MB XL;
2 Indigo2 R10k-195 512MB MaxImpact;
2 Indigo2 R4k-200 256MB (XL+Extreme);
2 Octane Dual R12K-300 1024MB (MXI+V6).
Hello,
I'm new in the SGI world (I bought an Octane2 after a lot of moths reading you :D ) and trying blender I have one problem:

When I start blender from the command line (./blender) the shell returns:
Code:
1121:./blender: rld: Error: unresolvable symbol in ./blender: llrint
1121:./blender: rld: Fatal Error: this executable has unresolvable symbols


Can you help me?

Thank you,
Alberto
PD: Octane 2 400Mhz Irix 6.5.22
Titox wrote:
Hello,
I'm new in the SGI world (I bought an Octane2 after a lot of moths reading you :D ) and trying blender I have one problem:

When I start blender from the command line (./blender) the shell returns:
Code:
1121:./blender: rld: Error: unresolvable symbol in ./blender: llrint
1121:./blender: rld: Fatal Error: this executable has unresolvable symbols


Can you help me?

Thank you,
Alberto
PD: Octane 2 400Mhz Irix 6.5.22


Alberto,
Welcome aboard. Read this topic: viewtopic.php?f=10&t=13198&p=103798&hilit=llrint#p103798

Two more things.

1. Start a new topic.
2. Use the search, you will find things like the above.


Regan

_________________
:Onyx2R: :Onyx2R: :0300: :0300: :0300: :O200: :Octane: :Octane: :O2: :O2: :Indigo2IMP: :Indy: :Indy: :Indy: :Indy: :Indy: :Indy: :Indy: :Indy:
:hpserv: J5600, 2 x SUN, 2 x Mac, 3 x Alpha, 2 x RS/6000
Tnx Regan, some packages of the overlays were not installed.

Quote:
Two more things.

1. Start a new topic.
2. Use the search, you will find things like the above.

Sorry, I'll do next time.

Alberto
Well, I need some help with compiling. I am coming across an instantiation loop that I cannot seem to get rid or, nor can I describe well enough the problem to the Blender group.

I have attached my Scons compile log, which includes all the errors.

I would really appreciate any help or insight!

Thanks,

_________________
Charles
Hi,

Does this help...?

C++ prelinker: error: instantiation loop in document:

http://www.techpubs.sgi.com/library/tpl ... %2B%2B_dev

Code:
7.13   Bug Fixes in MIPSpro 7.3.1.2m c++ fe


This section describes the C++ front-end bugs   that have been
fixed   since the MIPSpro 7.3.1.1m release.

755533    C++   7.3: /usr/lib32/cmplrs/fecc died due to   signal
4.

757655    CC Error: A   goto outside a region cannot jump to a
label inside a region.

760593    7.3   CC nm_elf with ksh/bsh and LD_LIBRARY[N32]
broken.

762310    Not   using defined operator new[] in   DT_SYMBOLIC
DSO   for class with virtual funcion.

768649    C++   Instantiation loop with   -Ofast.

770035    C++   prelinker: error in instantiation loop with
-IPA.



It probably builds better with gcc but run 200% slower.

Version of MIPS Pro are you using...?

Regan

_________________
:Onyx2R: :Onyx2R: :0300: :0300: :0300: :O200: :Octane: :Octane: :O2: :O2: :Indigo2IMP: :Indy: :Indy: :Indy: :Indy: :Indy: :Indy: :Indy: :Indy:
:hpserv: J5600, 2 x SUN, 2 x Mac, 3 x Alpha, 2 x RS/6000
Yeah, I already found those articles, thanks though.

I am running MipsPro 7.4.4 (I think, whatever the latest is), with all the current patches, and I am not using -Ofast or even -O3 (although I might try both to see if it changes how the problem pops up.)

_________________
Charles