LuxRender v0.7RC1 released

[SATtva]

Pinko, Linux-version has boost linked statically, it doesn't relies on system boost installation. What particular binary has the problem: luxconsole or luxrender? What about luxrender with Qt-static? Can you try to run Lux from console and post here any error messages it spits out?

[Henrik]

Aha, very lapidary .... no further comment than " i don´t care anymore"

Mh, sorry. I think it wasn't supposed to sound like that. I just wanted to point out that, much like the license agreement, there might be a good reason why it is there, but in most cases, there is no benefit for me in reading it, and I think I am not the only one with that. I did read the whole forum post about this release before I downloaded it, and it says something about Qt, but nothing Mac specific. I think that's the place where most people would expect important information like this.

[Sheltem]

Hi Guys,

LuxRender segfaults under ubuntu 8.1 (yeah a bit old I know) I get the same error for archive and static build...

Apr 5 19:43:56 Olive kernel: [ 5362.798241] luxrender[9413]: segfault at 1 ip b7295192 sp bfa02880 error 4 in libQtCore.so.4.4.3[b7149000+225000]

Which version of QT do I need to have installed? Do I read that right: 4.6.2?

Cheers!

[jensverwiebe]

Hi Guys,

LuxRender segfaults under ubuntu 8.1 (yeah a bit old I know) I get the same error for archive and static build...

Apr 5 19:43:56 Olive kernel: [ 5362.798241] luxrender[9413]: segfault at 1 ip b7295192 sp bfa02880 error 4 in libQtCore.so.4.4.3[b7149000+225000]

Which version of QT do I need to have installed? Do I read that right: 4.6.2?

Cheers!

Jupp, 4.6.2 or higher

Jens

[ayanematrix]

Ugh... maybe I'm the only one that uses a P.O.S. WinXP computer here, but it seems that this version of LuxRender eats almost twice as much memory for a given scene. v0.7RC1 eats over my 1.4GB of usable memory and crashes with a runtime error, whereas 0.6.1 (I think) and up till test builds in early March takes only 690MG for the same scene. Granted, they don't support the same features or bug fixes, but I'm confused on whether to file this as an actual bug or just an issue with my laptop.

If it is an issue on my computer, then there's not much more I can do with LuxRender until I can afford to get something much more robust.

[Lord Crc]

I just wanted to point out that, much like the license agreement, there might be a good reason why it is there, but in most cases, there is no benefit for me in reading it, and I think I am not the only one with that. I did read the whole forum post about this release before I downloaded it, and it says something about Qt, but nothing Mac specific. I think that's the place where most people would expect important information like this.

While I understand what you're saying, I find it slightly mind boggling. Though perhaps one can afford to have such a laid-back attitude on OSX. On Windows you'd end up with 6 million toolbars in your browser and 500 "helpful" programs bogging down your system.

[Lord Crc]

Ugh... maybe I'm the only one that uses a P.O.S. WinXP computer here, but it seems that this version of LuxRender eats almost twice as much memory for a given scene. v0.7RC1 eats over my 1.4GB of usable memory and crashes with a runtime error, whereas 0.6.1 (I think) and up till test builds in early March takes only 690MG for the same scene.

Just as the RC was out the door, we found (and fixed) a bug where imagemap textures would consume around 5x as much memory as before. I suspect this may be your issue. Could you try some geometry heavy scenes without textures?

[ayanematrix]

Ugh... maybe I'm the only one that uses a P.O.S. WinXP computer here, but it seems that this version of LuxRender eats almost twice as much memory for a given scene. v0.7RC1 eats over my 1.4GB of usable memory and crashes with a runtime error, whereas 0.6.1 (I think) and up till test builds in early March takes only 690MG for the same scene.

Just as the RC was out the door, we found (and fixed) a bug where imagemap textures would consume around 5x as much memory as before. I suspect this may be your issue. Could you try some geometry heavy scenes without textures?

5 times as much memory?! Wow, that's insane. I can assume that I'm getting hit with that as well to some degree as I see my textures loading up in the log window at least 3 times that of what I'm used to seeing; but, I can't tell if it's any worse as it's hitting the memory barrier I mentioned.

I'll see what I can pull up with the scene I was working on. Setting it to clay material only should remove the textures and make everything gray, right?

Edit: Okay, setting it to clay materials without any sort of textures allows for it load my scene, but it spams the following message a few times before it finishes the loading process:

Code: Select all

[2010-Apr-05 07:46:21 Warning: 43] Inconsistent shading normals

Also, memory usage is around 305.2 to 314.42MB actual and 344.64 to 353.82MB virtual with the last time I tried being almost 5 times as much before Lux went belly up. Ugh...

[Henrik]

While I understand what you're saying, I find it slightly mind boggling. Though perhaps one can afford to have such a laid-back attitude on OSX. On Windows you'd end up with 6 million toolbars in your browser and 500 "helpful" programs bogging down your system.

I don't have a lot experience with windows, but I think that the installation process is a bit more standardised in OS X. There is this general information window, and later on, you can choose what exactly you want to install. In case of LuxRender I can choose if I want LuxBlend installed. That's where you have to pay attention. And I don't remember a program that wanted to put a toolbar in my browser. I think they are not very common under OS X (or I'm just using the "wrong" software).

[Lord Crc]

Edit: Okay, setting it to clay materials without any sort of textures allows for it load my scene, but it spams the following message a few times before it finishes the loading process:

Code: Select all

[2010-Apr-05 07:46:21 Warning: 43] Inconsistent shading normals

Also, memory usage is around 305.2 to 314.42MB actual and 344.64 to 353.82MB virtual with the last time I tried being almost 5 times as much before Lux went belly up. Ugh...

The "inconsistent shading normals" message is triggered by "bad" vertex normals (ie pointing in when they should be pointing out). There was a slight bug in the detection code however for most meshes it's a genuine warning.

As for the memory usage it's no clear to me if the "clay" rendering showed normal memory usage or not.