A few days ago clement added extra things to normalize fireflies like that, the fixes in that report may be helping, altough the problem was focus on blend normals from the Ambient Oclussion, not sure if it is the same problem

I just tested with a0608340ae14 and i can't reproduce the error,
Win 7 - nvidia 1070 - Drv 430.86 - Wacom intuos 4 Drv:6.3.8

Hi, i have test with 4fb441517251, rendered that problematic shot a few times with different sizes and also other cameras, that blind guess seems to be working i didn't see any fire fly, this scene shots are rendering properly now.

1 open
2 render

I tested the fix with 8fa65ed31b7f and 8b2b79c2108b and the viewport looks great and mainly worked but. .

DANG! ok im feeling really really stupid right now, i though i had paked, and i was sure when was moving a file that weight more than an 1mb with only that little mesh data. . . . anyways, ill test this again in my secondary machine later to check this version.
Its kind a bit big but this is the file with the normal map.

Ok i have done more tests.
in my main machine (i7 win7 gtx 1070 ), updated drivers (430) an last blender version and the bug its consistently appearing always there.

ok, ill try to test this further, ill try to do it as soon as i can

This is what i get from the demo file.
Win 7 x64
Nvidia 1070 drivers 417.35

In T52685#458368, @Erick Tukuniata (erickblender) wrote:

"IMHO, here we'd rather need a new 'Copy Mirror' operator, that would properly duplicate and rename bones at the same time?"

Symmetrize is doing that already by it's symmetrizing it at the same time.

not sure, im not a developer, just a user that test and add info to save some dev time.

the operator that you are using there its "Pick shortest path" (bpy.ops.mesh.shortest_path_pick)
, not "select shortest path"( bpy.ops.mesh.shortest_path_select() )

Win 7 x64
I7 2600 16gb ram
nvidia 1070
Blender RC2 7b397cd

I can reproduce it in the 2.79 RC2 using a wacom intuos 4 stylus, mouse works fine.
win 7x64
nvidia 1070 drivers 382.33

(user testing here)
In both examples Fill region its active, that's causing that behavior.

@Dalai Felinto (dfelinto)
not sure if it helps now but i simplified the file isolating the the problematic face.

Im just theorizing here, but could you setup a render side by side like the last one, but setting the bevel to 0 in OSL or directly disabling it.
Again im not a developer, but my instinct in production its telling me that the difference in value its because of the lac of high lights cause by the bevel that's missing in the gpu side that not support OSL.

In T52625#456556, @LazyDodo (LazyDodo) wrote:

i have to be honest, not a developer im a user testing, downloaded the file started the render in cpu but didnt finish it i have to cancel, the render its huge and not sure where it is a transparent area. . .
That said.. . .
Tested on
I7 2600 16gb ram nvidia 1070 - Blender 2ca1f29
Render on cpu:
Start normal, didn't finish because of the long render time it would naturally a 2K image would take, so i canceled the render.
Render on Gpu:
Starts and renders and denoises, there is evident difference on the edges but couldn't notice anything particularly different.

tested using an nvidia 1070 on windows 7 x64 using Blender 2ca1f29
Here works as fine, no errors.

Blender version(build hash) of the build that its broken, a simple little file and step for reproduce the bug are required

Having the Bias value too low produce the artifacts, not sure if its intended or not but rising just a bit fix the issue.

could it be maybe this? developer.blender.org/T52341

Win 7 x64 blender build a8fbe99

Tested Confirm a8fbe99

And it seem that is because all the uv points are in the same place at (0 , 0)

confirm here: builds a8fbe99 6af7d7e

im not sure about several things i read here, probably need to study more im not the shiniest ball in the jar, so im not going to ask here since this is a bug track not a forum, just want to add some info about the ways that i use or work with this, maybe its a relevant info as a user point of view.

lets see if it is a bit more clear here
https://youtu.be/EAobUHJ5s30

i did more tests on a 3dsmax, importing from file import does not read the uvs, but importing object by object creating the primitive (alembic can be created as a primitive like a cube, click alembic on prim select the path the object click on the scene) it does read the uvs, also created rotated 90º but that its a max thing ^^
So it seem that yep the uvs are there, didn't try to export and import alembic on max to max, to check if its a soft thing or an alembic thing

also notice that to repeat the test from step one, at least in windows blender needs to be restarted or do a Ctrl+n to work a new file, directly reloading the file leaves a greyish mesh, not a white one as in step one.

Not a developer, but need to add more info,

i have a question about that, since the GP stroke sculpt can only be done under Stroke editing, is there a way to block other shortcuts that are unused? is not posible to select any uv island or move them or rotate them (G,R,S) anyways if the Gpencil Stroke editing its active
Other possibility( this is always as a user point of view) is to disable E as a quick short cut to sculpt or add a conditional key stroke at least in the uv image editor. Using pie menu D+E works fine, its not as practical as E but at least is there.

humm i wasn't aware of that.
If that's the case, sorry, my bad

ok
is this mailing list the place to raise the topic? ------> bf-translations-dev@blender.org ?

sorry i made a mistake on the version number, the last one i tested from builder that fails its
39ae324

Step to reproduce the error:
1-

to sum up, this termator thing affects reflections, refractions, bump maps or ¿can we call it shading distortion? and also shadows(actually shading not shadows per se)?
it seems to be a really big stuff not only a few things, basically affects every possible material that we could create exept volumetrics, emmit or non distorted shaders on high poly count models.
It could be possible to add some experimental feature at least to have a possibility to do a workaround using experimental stuff? or this will mess with some core functionality?
(yes i know this is not the place for feature request, im trying to understand how deep this problem is)

i didn't realized but yes i got the same on the consol
https://www.youtube.com/watch?v=1IxKx--qzc0&feature

~, ~ dyslexia and short sighted sorry, the working and double check build its Hash: 5c8fc8e

sorry the working version build was Hash: 5c8tc8e

after two more hours testing, i had reduce the sample file down to a single case

thanks for testing
i just downloaded both 32 and 64 version of the las build blender-2.74-3bffcc6 for windows, both still crash the same way, actually now also crash rendering using the button in the interface(but using the interface its a really random crash), so the deal must be among the object and the armature, removing the armature worked really well for me in my tests, maybe its on the material, i leave the material because removing it didn`t fully stabilized the render but made the crashes more random.

""Light Groups
In the Material buttons (F5) Shading panel, you can indicate a Group of Lamp Objects to be used as light for rendering, instead of the usual visible Lamps. This actually enables per-Material lighting (independent of which Object has the Material, or whether the Lamps are in visible layer).
If you don't want Lamps from a Light Group to work on the other Objects, just move them to an invisible layer.""

ok i fell a bit dumb because its weird but it works as it use to worked on blender 2.49 just cheked, sorry

just came up with a different situation but same bug,
on the step 1-fille above.
2-create a lamp
3-Ctrl+G to add the lamp to a group
4-F6 to access last operator options
5-change the name of the group
6-Accept the change using the key intro on the keyboard
7-switch to the original scene, just as the undo

Thanks
i imagine something like that could be happening, the work around was to copy the points from the vse the ones im interested on, to a new grease pencil data block and that work, but it took a wile to figure it out ^ -^,

the "Intro" submit the report before finish it.
2- hover on the right monitor (the vse) and select using A (that should work)
3- hit A again to deselect
4-try on the right monitor(the vse) to select a grease pencil point using Right click.
5- look at the crash.

this bug was reproduced under a linux system. . . y forgot to ask for info about that system, but at least i know its posible to repeat it outside windows