volumetric broken, not refresh #70091

Closed
opened 2019-09-20 04:13:03 +02:00 by noki paike · 22 comments

System Information
Operating system: Windows-10-10.0.17763 64 Bits
Graphics card: AMD Radeon HD 7600M Series ATI Technologies Inc. 4.5.13422 Core Profile Context 0

Blender Version
Broken: version: 2.81 (sub 12), branch: master, commit date: 2019-09-19 20:35, hash: c8df6f2cf9
Worked: (optional)

Short description of error
volumetric broken, not refresh, it is necessary to enter and exit in local mode to update the parameters

only on these gpus, I tried on the intel HD and they work well.

Exact steps for others to reproduce the error
see the gif
volumetrc.gif

**System Information** Operating system: Windows-10-10.0.17763 64 Bits Graphics card: AMD Radeon HD 7600M Series ATI Technologies Inc. 4.5.13422 Core Profile Context 0 **Blender Version** Broken: version: 2.81 (sub 12), branch: master, commit date: 2019-09-19 20:35, hash: `c8df6f2cf9` Worked: (optional) **Short description of error** volumetric broken, not refresh, it is necessary to enter and exit in local mode to update the parameters only on these gpus, I tried on the intel HD and they work well. **Exact steps for others to reproduce the error** see the gif ![volumetrc.gif](https://archive.blender.org/developer/F7759858/volumetrc.gif)
Author

Added subscriber: @nokipaike

Added subscriber: @nokipaike

#70308 was marked as duplicate of this issue

#70308 was marked as duplicate of this issue
Member

Added subscriber: @JacquesLucke

Added subscriber: @JacquesLucke
Member

Please provide a minimal example .blend file that allows us to reproduce the issue more easily.

Please provide a minimal example .blend file that allows us to reproduce the issue more easily.
Author

Added subscriber: @mano-wii

Added subscriber: @mano-wii
Author

here is the blend file, it's not really necessary, it happens with any scene, and with both volumetric shaders, it's a bug related to these GPUs.

volumetrics_bug.blend

for those who do a test, @mano-wii ^____^ it should disable the local auto zoom, otherwise is mess and you don't understand what happens.

Screenshot (174).png

here is the blend file, it's not really necessary, it happens with any scene, and with both volumetric shaders, it's a bug related to these GPUs. [volumetrics_bug.blend](https://archive.blender.org/developer/F7760597/volumetrics_bug.blend) for those who do a test, @mano-wii ^____^ it should disable the local auto zoom, otherwise is mess and you don't understand what happens. ![Screenshot (174).png](https://archive.blender.org/developer/F7760620/Screenshot__174_.png)
Author

I checked in blender 2.80 official it worked.

already didn't work before this:
3a08153d7a DRW: Refactor to support draw call batching

bug probably introduced by this:
caf494444c EEVEE: Volume: Speedup: Use Image load store extension

I checked in blender 2.80 official it worked. already didn't work before this: 3a08153d7a84 DRW: Refactor to support draw call batching bug probably introduced by this: caf494444cd7 EEVEE: Volume: Speedup: Use Image load store extension
Author

I made a comparison video between blender 2.80 and the latest build of 2.81
the samples loop refresh appears to be broken
(the refresh of the first samples, in the video it does it because it is I who press the switch between local and normal view)

volumetric.mp4

I made a comparison video between blender 2.80 and the latest build of 2.81 the samples loop refresh appears to be broken (the refresh of the first samples, in the video it does it because it is I who press the switch between local and normal view) [volumetric.mp4](https://archive.blender.org/developer/F7771956/volumetric.mp4)
Member

Added subscribers: @fclem, @lichtwerk

Added subscribers: @fclem, @lichtwerk
Member

CC @fclem

CC @fclem
Author

This comment was removed by @nokipaike

*This comment was removed by @nokipaike*
Author

I did tests with a nebula file a little more complex, it always has the problem of refreshing, but it seems to work partially.

volume_so_better.mp4

Nebula_ColorNodes.blend

I did tests with a nebula file a little more complex, it always has the problem of refreshing, but it seems to work partially. [volume_so_better.mp4](https://archive.blender.org/developer/F7779093/volume_so_better.mp4) [Nebula_ColorNodes.blend](https://archive.blender.org/developer/F7779092/Nebula_ColorNodes.blend)

Added subscriber: @brecht

Added subscriber: @brecht

Changed status from 'Open' to: 'Archived'

Changed status from 'Open' to: 'Archived'
Brecht Van Lommel self-assigned this 2019-10-02 23:44:29 +02:00

This GPU is not officially supported, so I think this can be closed. Not supported really means that, we should not spend time on it but focus on making the supported GPUs work well.

This GPU is not officially supported, so I think this can be closed. Not supported really means that, we should not spend time on it but focus on making the supported GPUs work well.
Author

@brecht
but come on, what way is this to solve problems,

the volumetrics worked well in this gpu until a few weeks ago of blender 2.81!

close a bug report without even giving it a look at the code
maybe you find out that it is a problem that takes two minutes.

In #70091#788179, @brecht wrote:
This GPU is not officially supported, so I think this can be closed. Not supported really means that, we should not spend time on it but focus on making the supported GPUs work well.

also means that they should not be broken ... don't you think?
if the volumetric had never worked, I could also understand it.

And I'm sure some newer AMD gpu will come out with broken volumetrics more recently ......
Still not many have done extensive checks on volumetric on blender 2.81

@brecht but come on, what way is this to solve problems, the volumetrics worked well in this gpu until a few weeks ago of blender 2.81! close a bug report without even giving it a look at the code maybe you find out that it is a problem that takes two minutes. > In #70091#788179, @brecht wrote: > This GPU is not officially supported, so I think this can be closed. Not supported really means that, we should not spend time on it but focus on making the supported GPUs work well. also means that they should not be broken ... don't you think? if the volumetric had never worked, I could also understand it. And I'm sure some newer AMD gpu will come out with broken volumetrics more recently ...... Still not many have done extensive checks on volumetric on blender 2.81

also means that they should not be broken ... don't you think?

No, unsupported means we don't spend time on it and things should not be expected to work.

> also means that they should not be broken ... don't you think? No, unsupported means we don't spend time on it and things should not be expected to work.
Author

They really worked well. especially volumetric.
They worked in blender 2.80, they worked until a couple of weeks ago. Until then, EVERYTHING worked in this gpu.

Of course, workarounds were needed, and now they exist and these work. but you have decided that these GPUs must now work partially.

But oh well, who cares, for a few million users who own these generations of GPUs, they will have to be thrown away. We all order a new GPU. Or rather specify a new laptop in my case if I want to use blender in a worthy way.
Now that you will do when so many other people, especially after blender 2.81 is officially released, and in this gpu everything worked in blender 2.80, will you torpedo them immediately and continuously because this million gpus no longer has a right to a workaround or a bug that is fixed?

You haven't even seen what it this bug, is and you have closed the report.
it's just a refresh loop that doesn't work anymore for some reason.
I even specified the right time when it stopped working.

You have decided to snub it.
Not even low priority bug, directly archiveid.

I apologize, but I'm pissed, put yourself in my shoes ...

They really worked well. especially volumetric. They worked in blender 2.80, they worked until a couple of weeks ago. Until then, EVERYTHING worked in this gpu. Of course, workarounds were needed, and now they exist and these work. but you have decided that these GPUs must now work partially. But oh well, who cares, for a few million users who own these generations of GPUs, they will have to be thrown away. We all order a new GPU. Or rather specify a new laptop in my case if I want to use blender in a worthy way. Now that you will do when so many other people, especially after blender 2.81 is officially released, and in this gpu everything worked in blender 2.80, will you torpedo them immediately and continuously because this million gpus no longer has a right to a workaround or a bug that is fixed? You haven't even seen what it this bug, is and you have closed the report. it's just a refresh loop that doesn't work anymore for some reason. I even specified the right time when it stopped working. You have decided to snub it. Not even low priority bug, directly archiveid. I apologize, but I'm pissed, put yourself in my shoes ...
Author

@mano-wii sorry if I insist on this, but I think it's not a problem related to volumetrics system ... and I'll explain why ..
Remember when the new eevee batch cache system was implemented, and that the framebuffer didn't update the mesh position on these GPUs? ... it was this bug #69853 which has been successfully solved by you ...

I believe that it is the same problem, but linked to the shading system .. look at this video ...

just click on a new position of the 3d cursor and everything "re-works" but only for that framebuffer, then if I move the objects don't refresh anymore until I reposition the 3d cursor.

refresh.mp4

fishy.mp4

why do we need to position the 3d cursor so that the framebuffer is refreshed?

I think it's due to the fact that this problem is related to the old workaround that is still partially working for some shaders and creates these refresh problems...

so what happens if the old workaround that no longer makes sense to exist were removed altogether?
i think maybe could solve this problem...


this is the nebula file in case you want to do a quick test by simply "turning off the old workaround."
Nebula_ColorNodes.blend

@mano-wii sorry if I insist on this, but I think it's not a problem related to volumetrics system ... and I'll explain why .. Remember when the new eevee batch cache system was implemented, and that the framebuffer didn't update the mesh position on these GPUs? ... it was this bug #69853 **which has been successfully solved by you ...** I believe that it is the same problem, but linked to the shading system .. look at this video ... just click on a new position of the 3d cursor and everything "re-works" but only for that framebuffer, then if I move the objects don't refresh anymore until I reposition the 3d cursor. [refresh.mp4](https://archive.blender.org/developer/F7845067/refresh.mp4) [fishy.mp4](https://archive.blender.org/developer/F7845062/fishy.mp4) why do we need to position the 3d cursor so that the framebuffer is refreshed? I think it's due to the fact that this problem is related to the old workaround that is still partially working for some shaders and creates these refresh problems... so what happens if the old workaround that no longer makes sense to exist were removed altogether? i think maybe could solve this problem... ____ this is the nebula file in case you want to do a quick test by simply "turning off the old workaround." [Nebula_ColorNodes.blend](https://archive.blender.org/developer/F7845075/Nebula_ColorNodes.blend)

I'm not sure where the problem is, but I know this change solves it:

diff --git a/source/blender/draw/engines/eevee/eevee_volumes.c b/source/blender/draw/engines/eevee/eevee_volumes.c
index aaa351a1922..8220c0e9501 100644
--- a/source/blender/draw/engines/eevee/eevee_volumes.c
+++ b/source/blender/draw/engines/eevee/eevee_volumes.c
@@ -41,6 +41,7 @@
 
 - include "eevee_private.h"
 - include "GPU_draw.h"
+#include "GPU_extensions.h"
 - include "GPU_texture.h"
 - include "GPU_material.h"
 
@@ -81,7 +82,9 @@ extern char datatoc_volumetric_integration_frag_glsl[];
 extern char datatoc_volumetric_lib_glsl[];
 extern char datatoc_common_fullscreen_vert_glsl[];
 
-#define USE_VOLUME_OPTI (GLEW_ARB_shader_image_load_store && GLEW_ARB_shading_language_420pack)
+#define USE_VOLUME_OPTI \
+  (GLEW_ARB_shader_image_load_store && GLEW_ARB_shading_language_420pack && \
+   !GPU_crappy_amd_driver())
 
 static void eevee_create_shader_volumes(void)
 {

I'm not sure where the problem is, but I know this change solves it: ``` diff --git a/source/blender/draw/engines/eevee/eevee_volumes.c b/source/blender/draw/engines/eevee/eevee_volumes.c index aaa351a1922..8220c0e9501 100644 --- a/source/blender/draw/engines/eevee/eevee_volumes.c +++ b/source/blender/draw/engines/eevee/eevee_volumes.c @@ -41,6 +41,7 @@ - include "eevee_private.h" - include "GPU_draw.h" +#include "GPU_extensions.h" - include "GPU_texture.h" - include "GPU_material.h" @@ -81,7 +82,9 @@ extern char datatoc_volumetric_integration_frag_glsl[]; extern char datatoc_volumetric_lib_glsl[]; extern char datatoc_common_fullscreen_vert_glsl[]; -#define USE_VOLUME_OPTI (GLEW_ARB_shader_image_load_store && GLEW_ARB_shading_language_420pack) +#define USE_VOLUME_OPTI \ + (GLEW_ARB_shader_image_load_store && GLEW_ARB_shading_language_420pack && \ + !GPU_crappy_amd_driver()) static void eevee_create_shader_volumes(void) { ```

This issue was referenced by d0d57f1712

This issue was referenced by d0d57f17121f1649b4347a14769d274204945717

Changed status from 'Archived' to: 'Resolved'

Changed status from 'Archived' to: 'Resolved'
Sign in to join this conversation.
No Label
Interest
Alembic
Interest
Animation & Rigging
Interest
Asset Browser
Interest
Asset Browser Project Overview
Interest
Audio
Interest
Automated Testing
Interest
Blender Asset Bundle
Interest
BlendFile
Interest
Collada
Interest
Compatibility
Interest
Compositing
Interest
Core
Interest
Cycles
Interest
Dependency Graph
Interest
Development Management
Interest
EEVEE
Interest
EEVEE & Viewport
Interest
Freestyle
Interest
Geometry Nodes
Interest
Grease Pencil
Interest
ID Management
Interest
Images & Movies
Interest
Import Export
Interest
Line Art
Interest
Masking
Interest
Metal
Interest
Modeling
Interest
Modifiers
Interest
Motion Tracking
Interest
Nodes & Physics
Interest
OpenGL
Interest
Overlay
Interest
Overrides
Interest
Performance
Interest
Physics
Interest
Pipeline, Assets & IO
Interest
Platforms, Builds & Tests
Interest
Python API
Interest
Render & Cycles
Interest
Render Pipeline
Interest
Sculpt, Paint & Texture
Interest
Text Editor
Interest
Translations
Interest
Triaging
Interest
Undo
Interest
USD
Interest
User Interface
Interest
UV Editing
Interest
VFX & Video
Interest
Video Sequencer
Interest
Virtual Reality
Interest
Vulkan
Interest
Wayland
Interest
Workbench
Interest: X11
Legacy
Blender 2.8 Project
Legacy
Milestone 1: Basic, Local Asset Browser
Legacy
OpenGL Error
Meta
Good First Issue
Meta
Papercut
Meta
Retrospective
Meta
Security
Module
Animation & Rigging
Module
Core
Module
Development Management
Module
EEVEE & Viewport
Module
Grease Pencil
Module
Modeling
Module
Nodes & Physics
Module
Pipeline, Assets & IO
Module
Platforms, Builds & Tests
Module
Python API
Module
Render & Cycles
Module
Sculpt, Paint & Texture
Module
Triaging
Module
User Interface
Module
VFX & Video
Platform
FreeBSD
Platform
Linux
Platform
macOS
Platform
Windows
Priority
High
Priority
Low
Priority
Normal
Priority
Unbreak Now!
Status
Archived
Status
Confirmed
Status
Duplicate
Status
Needs Info from Developers
Status
Needs Information from User
Status
Needs Triage
Status
Resolved
Type
Bug
Type
Design
Type
Known Issue
Type
Patch
Type
Report
Type
To Do
No Milestone
No project
No Assignees
7 Participants
Notifications
Due Date
The due date is invalid or out of range. Please use the format 'yyyy-mm-dd'.

No due date set.

Dependencies

No dependencies set.

Reference: blender/blender#70091
No description provided.