Author Topic: Questions to understand about nouveau problem with NV4x chipsets  (Read 2007 times)

0 Members and 1 Guest are viewing this topic.

Questions to understand about nouveau problem with NV4x chipsets
Hi,

I found a problem with NV4x chipsets (GeForce 7300) and SW OT mod, on Ubuntu 16.04 with oibaf mesa git;

the problem is that some models produce full screen z-culling shaders artifact (resulting in X-Wing model in Tech Room occupying all screen an overlapping to game menu/GUI, which disappears completely when selecting Tie-Fighter or Tie-Interceptor)

I would like to ask if someone encounterd similar problems on other games/mods with the same NV4x chipsets.

Following indications of mesa developer Ilia Mirkin and Issue on github talking about the need of setting #version 120 header,
I tried that but the result is still the same X-Wing shaders covering all screen in dark artifacts.

The qapitrace showing artifact is at the following link, but just people with Ubuntu and NV4x with gallium nouveau driver would see the artifacts with apitrace-gui retracer:

https://drive.google.com/open?id=0B_OFHiIqgpSFUUtnTXdWSEhoRnc

Now, the next thing to investigate is the following, a question proposed by Ilia, could you help me to check about combination of color and depth buffers in use by fs2_open (3.7.x and previous versions are all affected with NV4x)?

Quote
If you are involved with the development of this application, could you verify that it never uses a 32-bit color buffer with a 16-bit depth buffer, or a 16-bit color buffer with a 24-bit depth buffer? (That's RGBA8 + Z16 or RGB565 + Z24.) Those combinations, while perfectly legal in GL, are not actually supported by the hardware, and the driver "works around it" by just disabling depth, which results in broken rendering (but at least no hangs...).

If that's the case and the driver works around we may argue that it's not doing its best,
but do you know if there may be a way to enhance compatibility in fs2_open?

Or do you know a way to apply to X-Wing model some changes to shaders in order to avoid the problem (as majority of the models in SW OT do not cause any problem)?

Thanks for your attention and feedbacks
Even if NV4x may seem obsolete there are still some fellows playing with that.

Mauro
« Last Edit: June 01, 2016, 05:59:39 pm by maurossi »

 

Offline m!m

  • 211
Re: Questions to understand about nouveau problem with NV4x chipsets
Have you looked at the fs2_open log file? That should contain the bit size values reported by SDL. The default color window buffers should always either use 16 bit color and 16 bit depth or 24/32 bit color and 24 bit depth buffers.

 
Re: Questions to understand about nouveau problem with NV4x chipsets
Hi,

The color depths seem not causing the issue, which affects specifically x-wing and a-wing models.

Adding new information acquired recently,

if I load the x-wing and a-wing models in PCS 2.1 on Ubuntu (latest wxwidgets 3.0 one, with OpenGL VBO enabled)
on NV4x chipsets there are no culling artifacts,

while on fs2_open_3.7.2 and fs_open_3.7.4 the culling artifacts are still there with x-wing and a-wing.

Do you know what differences are there between fs2_open and PCS 2.1 rendering engines that may explain the culling artifacts?

Mauro

 

Offline DahBlount

  • Moderator
  • 29
  • Alpine ☆ Cancer Tribulation
    • Minecraft
    • Skype
    • Steam
Re: Questions to understand about nouveau problem with NV4x chipsets
What about recent nightly builds (3.7.5)? 3.7.4 is pre-OpenGL Core support(as well as a bunch of other OGL related refectors) in FSO, and I believe there were issues with the Mesa drivers and FSO before OGL Core support got merged in.

Current PCS2 initializes OGL much like post-OGL Core FSO does and there were issues with culling artifacts in the WxW 3.0.2 pull requests for PCS2 before we solved the OGL Init issue.
<Axem> yet still more insightful than #hard-light

<Axem> jad2.23 will just be cat videos

<DahBlount> So
<DahBlount> JAD2.2 is like that
<Axem> maybe
<Axem> it can be whatever you like!
<DahBlount> A Chocolate Sundae?
<Axem> sure

My models: GTF Gilgamesh - GTD Nuadha [Redesigning] - Ningirama [WIP] - GTG Zephyrus

 

Offline The E

  • He's Ebeneezer Goode
  • Moderator
  • 213
  • Nothing personal, just tech support.
    • Steam
    • Twitter
Re: Questions to understand about nouveau problem with NV4x chipsets
What about recent nightly builds (3.7.5)? 3.7.4 is pre-OpenGL Core support(as well as a bunch of other OGL related refectors) in FSO, and I believe there were issues with the Mesa drivers and FSO before OGL Core support got merged in.

Current PCS2 initializes OGL much like post-OGL Core FSO does and there were issues with culling artifacts in the WxW 3.0.2 pull requests for PCS2 before we solved the OGL Init issue.

If this is really about an NV4x-GPU, then that won't help. Those GPUs (which were used in the 6k and low-end 7k series) do not offer OpenGL 3.0 support, which would be needed to run the OpenGL Core graphics code.

Ultimately, and I am very sorry to say this, our ability to support 10 year old entry-level GPUs is somewhat limited.
If I'm just aching this can't go on
I came from chasing dreams to feel alone
There must be changes, miss to feel strong
I really need lifе to touch me
--Evergrey, Where August Mourns