[Mesa-dev] [Bug 94341] Incorrect results with dFdx / dFdy

2016-03-01 Thread bugzilla-daemon
https://bugs.freedesktop.org/show_bug.cgi?id=94341

Ian Kerr  changed:

   What|Removed |Added

 Status|NEW |RESOLVED
 Resolution|--- |NOTABUG

--- Comment #2 from Ian Kerr  ---
Thanks Brian! Yes, I was using the old "swrast". Recompiling osmesa with
llvmpipe support seems to have fixed the issue.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev


[Mesa-dev] [Bug 94341] Incorrect results with dFdx / dFdy

2016-02-29 Thread bugzilla-daemon
https://bugs.freedesktop.org/show_bug.cgi?id=94341

--- Comment #1 from Brian Paul  ---
Are you using the old "swrast" driver?  If so, the derivative functions only
work correctly for some values (the interpolated texcoords, colors, etc.).

If you can switch to the llvmpipe driver you shouldn't have that problem (and
your code should run faster).

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev


[Mesa-dev] [Bug 94341] Incorrect results with dFdx / dFdy

2016-02-29 Thread bugzilla-daemon
https://bugs.freedesktop.org/show_bug.cgi?id=94341

Roland Scheidegger  changed:

   What|Removed |Added

 Attachment #122036|text/plain  |image/png
  mime type||

-- 
You are receiving this mail because:
You are the assignee for the bug.
You are the QA Contact for the bug.___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev


[Mesa-dev] [Bug 94341] Incorrect results with dFdx / dFdy

2016-02-29 Thread bugzilla-daemon
https://bugs.freedesktop.org/show_bug.cgi?id=94341

Bug ID: 94341
   Summary: Incorrect results with dFdx / dFdy
   Product: Mesa
   Version: 11.1
  Hardware: Other
OS: Linux (All)
Status: NEW
  Severity: normal
  Priority: medium
 Component: Mesa core
  Assignee: mesa-dev@lists.freedesktop.org
  Reporter: ik...@box.com
QA Contact: mesa-dev@lists.freedesktop.org

Created attachment 122036
  --> https://bugs.freedesktop.org/attachment.cgi?id=122036=edit
Correct result (left) and incorrect result (right)

Context: I am using Chrome and osmesa to perform "headless" WebGL rendering.
Many of the Three.js examples (http://threejs.org/examples/) do not render
correctly. I think I have narrowed the problem down to incorrect results
produced by dFdx / dFdy. Three.js uses these functions for several things
including: flat-shading and bump / normal mapping.

I have created a JSFiddle that demonstrates the problem:

https://jsfiddle.net/iangkerr/08x72ge9/

The fiddle renders two triangles, rotated 45 degrees, with a perspective
projection. The fragment shader attempts to compute the surface normal and
assigns the result to gl_FragColor. I expect a purplish colour, due to the
normal having positive X and Z components (R and B), and I get this result on
my Mac. However, when running Chrome + osmesa on Ubuntu, I get a bluish colour
(positive Z component). I have attached a screenshot of the correct result and
the incorrect result.

-- 
You are receiving this mail because:
You are the QA Contact for the bug.
You are the assignee for the bug.___
mesa-dev mailing list
mesa-dev@lists.freedesktop.org
https://lists.freedesktop.org/mailman/listinfo/mesa-dev