john peterson (bliblubli)
User

Projects

User does not belong to any projects.
User Since
Jun 2 2013, 4:47 PM (194 w, 15 h)

Recent Activity

Fri, Feb 17

john peterson (bliblubli) awarded D2513: Cycles: Speed up by avoiding extra calulations when unneeded a Like token.
Fri, Feb 17, 6:37 PM

Thu, Feb 16

john peterson (bliblubli) created T50703: numpy crashes on latest linux buildbot.
Thu, Feb 16, 7:30 PM · Python, BF Blender

Fri, Feb 10

john peterson (bliblubli) added a comment to T50473: inset and extrude sometime broken.

This object with this mesh "happened" during work, without script. Sadly, I just noticed the weird behavior to late. The last steps I could remember I did, didn't permit to recreate the bug.
Applying the scale indeed resolves the problem.
But still, why does the extrude properly work on 2.74 and not in master?
Well, I guess it's not a that problematic if this bug stays in master, it seems to only happen very rarely anyway. and 2.74 allowing to extrude an object in a direction that is scaled to 0 is non-sens anyway.
As the steps to lead to that mesh are nearly impossible to find, we can just let it archived until someone get the same problem.

Fri, Feb 10, 1:49 PM · Mesh Modeling, BF Blender

Thu, Jan 26

john peterson (bliblubli) added a comment to T50461: Some modifiers make render crashes right away on complex files.

I managed to recreate files that crash on render start with --factory-startup with dummy meshes and textures. But like @matali23 (matali23), when I reopen the file and hit render, it works. So submitting the file wouldn't help much. But it's a real problem, so it's sad if it get's ignored.

Thu, Jan 26, 6:41 PM · BF Blender

Jan 19 2017

john peterson (bliblubli) updated subscribers of T50461: Some modifiers make render crashes right away on complex files.

@Joey Ferwerda (TheOnlyJoey) I report bugs since 3 years and a half now. I must say this bug is really nasty. I already tried to reproduce it by replacing every mesh with a subdivided cube with similar polycount and every texture with a generated one of same resolution. But even without changing polycount, object count and texture count and sizes, the file just wouldn't crash anymore.
As I said in my comment, most of the time, only reopening the same file, without changing anything is enough to workaround the bug. So I don't see how a file could be submitted here. Maybe let @Bastien Montagne (mont29) and @Sergey Sharybin (sergey) decide if the crash log is enough to find the bug?

Jan 19 2017, 6:39 PM · BF Blender
john peterson (bliblubli) created T50473: inset and extrude sometime broken.
Jan 19 2017, 1:38 PM · Mesh Modeling, BF Blender

Jan 18 2017

john peterson (bliblubli) assigned T50276: intersect fails sometime depending on the shapekey to Howard Trickey (howardt).
Jan 18 2017, 9:18 PM · BF Blender
john peterson (bliblubli) updated subscribers of T50276: intersect fails sometime depending on the shapekey.

After discussion with @Luca Rood (LucaRood) assigning to @Howard Trickey (howardt).

Jan 18 2017, 9:18 PM · BF Blender
john peterson (bliblubli) added a comment to T50461: Some modifiers make render crashes right away on complex files.

Maybe related to this https://developer.blender.org/T48834#410518. Anyway, I can confirm that I experience exact same problem from time to time on big scenes (crash at render begin or when render finishes.) But it's rare enough (about once a day) to be acceptable in my case, although it's annoying to lose 30min of rendering. But restarting blender always worked for me.

Jan 18 2017, 7:47 PM · BF Blender

Jan 16 2017

john peterson (bliblubli) added a comment to D2383: Cycles: Add option to replace GI with AO approximation after certain amount of bounces.

@Sergey Sharybin (sergey) Ok, great if it's intended to work on GPU. For some reason, it doesn't work with the split kernel. Everything compiles and renders, but without any replacement by GI after the chosen bounce number. On CPU, it works great.

Jan 16 2017, 9:01 PM

Jan 14 2017

john peterson (bliblubli) added a comment to D2383: Cycles: Add option to replace GI with AO approximation after certain amount of bounces.

Great speedups on CPU indeed. Is there a limitation making it impossible/not worth on GPU?

Jan 14 2017, 9:56 PM
john peterson (bliblubli) added a comment to D2348: Cycles: Refactor split kernel and implement for CPU.

If some improvement visible to user (in speed or functionnality like SSS and volumetrics) can be reached for 2.79 with this new code, I would add it to master.
If benefits are only for programmers, I would add it to 2.8. Otherwise, it would just be risking breaking rendering on different configurations, cluttering the bug tracker more than it is already and maybe releasing a 2.79 (which may be used over a long period by many users) that will not be stable.

Jan 14 2017, 2:21 PM

Jan 10 2017

john peterson (bliblubli) triaged T50103: Parented object with Scale 1,1,0 seems to have Loc and Rot locked. as "Confirmed" priority.

Can confirm the bug. Indeed, changing the rotation value in the N-Panel works, but the modal rotation in the 3D view doesn't work anymore. Having a scale not equal to 0 deactivate the bug.

Jan 10 2017, 10:00 PM · BF Blender
john peterson (bliblubli) triaged T50276: intersect fails sometime depending on the shapekey as "Confirmed" priority.

Can confirm the bug for both files.

Jan 10 2017, 9:55 PM · BF Blender
john peterson (bliblubli) triaged T50162: Selection from big data sets as "Normal" priority.

Indeed, selection is wrong, can confirm, but it's a known limitation that Blender can't handle large object counts.
I'll however let the bug open until it's sure that the problem really comes from the amount of object.

Jan 10 2017, 9:54 PM · BF Blender
john peterson (bliblubli) assigned T50396: pdf geometry is sometime rotated 180° to Howard Trickey (howardt).

I can confirm the bug.

Jan 10 2017, 9:50 PM · Import/Export, Addons
john peterson (bliblubli) triaged T50166: Cycles - Viewport render preview stucked on the first sample - in a scene a DupliVerts objet parenting instances owned by more than one Groups as "Confirmed" priority.

Can confirm the bug, viewport render restarts indefinetly.

Jan 10 2017, 9:46 PM · BF Blender, Cycles

Jan 7 2017

john peterson (bliblubli) added a comment to T50333: Blender closes when using texture paint in 3d view.

Can't reproduce the bug here on Linux, painting in UV/image editor and clicking many times in the 3D view.
Looks like a driver issue. Windows 10 is known to be pretty unstable also, installing new drivers that are rather experimental without asking. Attach your system info (help menu -> save system info).

Jan 7 2017, 4:56 PM · Sculpting and Painting, BF Blender
john peterson (bliblubli) assigned T50338: Bevel + Subsurf + Edge Crease = jagged edges to Howard Trickey (howardt).

Can confirm the behavior. Really looks like a bug indeed. It only happens when the bevel modifier is activated, so I'll let @Howard Trickey (howardt) decide.

Jan 7 2017, 4:50 PM · Mesh Modeling, BF Blender
john peterson (bliblubli) updated subscribers of T50382: Fluid sim broken when domain/fluid are at large coordinates.
Jan 7 2017, 4:28 PM · BF Blender
john peterson (bliblubli) triaged T50344: Bug Report: Instance Modifier as "Incomplete" priority.

The size cannot be changed, because of the driver it has. Deleting the driver and increasing the size doesn't show up more particles. By the way, the size is in particle system/ physics section, not render. There is a scale option in render, but deactivating or activating it doesn't change anything, so I guess it's not what you meant either?
You worked hard for this report, but I can't see what is not working as expected. Please add a screen-shot of how it looks and how you think it should look.

Jan 7 2017, 4:23 PM · Modifiers, BF Blender
john peterson (bliblubli) reopened T50382: Fluid sim broken when domain/fluid are at large coordinates as "Open".
Jan 7 2017, 4:12 PM · BF Blender
john peterson (bliblubli) added a comment to T50370: Problems (yet again) with Linked Objects that have Alembic Cache applied.

@Ajlan Altug (jacobo) can't reproduce issue, but it may be due to missing alembic file?

Jan 7 2017, 4:02 PM · BF Blender
john peterson (bliblubli) closed T50382: Fluid sim broken when domain/fluid are at large coordinates as "Resolved".

Can confirm the bug. Is pretty annoying as it can't even be canceled. One has to kill the process to cancel it.

Jan 7 2017, 3:53 PM · BF Blender

Jan 3 2017

john peterson (bliblubli) added a comment to D2280: Add missing UV tools in the UV/Image Editor Window > Tools tab.

@Aaron Carlisle (Blendify) put some reviewers to get it reviewed, looks good.

Jan 3 2017, 3:21 PM · User Interface, BF Blender

Dec 22 2016

john peterson (bliblubli) added a comment to T50277: blender crushes periodically.

@Aaron Carlisle (Blendify) I must say I also experience a lot of non-reproducible crashes since 3-4 weeks with buildbots. I set the auto save at 1 min to be able to restore the file before it crashed, but it doesn't help. Redoing the exact same steps after opening the autosave doesn't trigger a crash. It happens most of the time with material and when entering edit mode, but also sometime at rendering start or end.
Really don't how to report those bugs, but they are there.

Dec 22 2016, 5:59 PM · BF Blender

Dec 17 2016

john peterson (bliblubli) added a comment to T50181: Cycles-Regression: invisible shaders are optimized and compiled although they shouldn't.
Dec 17 2016, 9:13 AM · Cycles, BF Blender

Dec 16 2016

john peterson (bliblubli) added a comment to D113: Datablock ID Properties.

From a design point of view, will it allow to link/append node trees from addons (like Luxblend for Luxcore or animation nodes) with needed textures, meshes and whatever can be referenced in a node tree?

Dec 16 2016, 6:27 PM

Dec 8 2016

john peterson (bliblubli) added a comment to D2406: Cycles: Experiment with cutout closure.

That new effect is really cool :) I can definitely see some good usage for it in engineering and architecture. And it then makes sense to have it as a node.

Dec 8 2016, 9:58 PM

Dec 7 2016

john peterson (bliblubli) added a comment to T50181: Cycles-Regression: invisible shaders are optimized and compiled although they shouldn't.

Tried different revisions, will try more tomorrow, but at least in may 2015, it worked. So it happened after may 2015. Just one year and a half of commits to check :D

Dec 7 2016, 7:48 PM · Cycles, BF Blender

Dec 4 2016

john peterson (bliblubli) added a project to T50181: Cycles-Regression: invisible shaders are optimized and compiled although they shouldn't: Cycles.
Dec 4 2016, 5:53 PM · Cycles, BF Blender
john peterson (bliblubli) updated subscribers of T50181: Cycles-Regression: invisible shaders are optimized and compiled although they shouldn't.
Dec 4 2016, 5:53 PM · Cycles, BF Blender
john peterson (bliblubli) created T50181: Cycles-Regression: invisible shaders are optimized and compiled although they shouldn't.
Dec 4 2016, 5:53 PM · Cycles, BF Blender

Nov 20 2016

john peterson (bliblubli) added a comment to T49991: reloading librairies doesn't update node groups.

@Bastien Montagne (mont29) thanks for the fix. It is also my feeling that Linux Builds crashes much less. At least in this case, windows help get a cleaner code :)

Nov 20 2016, 7:12 PM · BF Blender

Nov 15 2016

john peterson (bliblubli) updated the summary for D2339: First accepted part of D2254.
Nov 15 2016, 7:55 PM

Nov 14 2016

john peterson (bliblubli) updated the diff for D2339: First accepted part of D2254.

added a check for the integrator setting "cycles.use_transparent_shadows" as user may disable transparent shadows even if some transparent shaders are present in the scene. Thanks to @Mai Lavelle (maiself) for her help on finding this solution.

Nov 14 2016, 10:54 PM
john peterson (bliblubli) added a comment to D2341: split-kernel: add new node features to speedup heavy production files..

We've talked to AMD about AoS/SoA and big switch statements. The feedback so far is that SoA/AoS should not really matter with their latest drivers. Assuming that eventually driver will be improved on Linux i think we should stick to SoA topology. Big switch statement is still an open question tho, so guess we'll have to accept such black magic with re-ordering for the time being.

Nov 14 2016, 9:42 PM
john peterson (bliblubli) updated the summary for D2341: split-kernel: add new node features to speedup heavy production files..
Nov 14 2016, 9:39 PM
john peterson (bliblubli) updated the summary for D2341: split-kernel: add new node features to speedup heavy production files..
Nov 14 2016, 9:39 PM
john peterson (bliblubli) updated the diff for D2341: split-kernel: add new node features to speedup heavy production files..

Tab to spaces and some wrapping corrected. For some reason, the wireframe line still is not aligned...

Nov 14 2016, 9:38 PM
john peterson (bliblubli) added a comment to D2340: Split-kernel: reorganize nodes to get maximal performance on Group 0 (clay render) and better perfs on group 1 and 2..

@Sergey Sharybin (sergey) Well I thought the reorganisation alone didn't give any speedup, which is true for BMW. But I've tested other files and got following result on Win7 x64 with 16.11.2:

ScenesppPreparation timemasterD2340D2341gain (over master) reorganisationgain (over master) selective node compilation
Barcelona6400:00.7601:08.7001:07.01.8401:03.05-3%-8%
Fishy Cat25600:08.9202:13.4002:12.9302:12.91-0.5%-0.5%
Koro6400:11.3003:46.1903:44.4403:42.13-1%-2%

Interesting to note that gains on Windows are comparable to those on Linux for Barcelona at least with -8% with D2341, although Linux is way faster than Windows again.

Nov 14 2016, 7:45 PM · Cycles
john peterson (bliblubli) updated the summary for D2339: First accepted part of D2254.
Nov 14 2016, 7:20 PM

Nov 13 2016

john peterson (bliblubli) added a comment to T49915: dxf import fails for white surfaces and rotates blocks 90°.

I can confirm the bug and it's pretty annoying. Looking back at old projects, I did some serious mistakes because of this 90° rotation only happening on blocks. Would be great to have the fix backported to 2.78.

Nov 13 2016, 10:54 AM · Import/Export, Addons
john peterson (bliblubli) added a comment to D2341: split-kernel: add new node features to speedup heavy production files..

@Mai Lavelle (maiself): first bunch of benchmark with this patch, I'll be updating as it comes:
Ubuntu16.04 with AMDGPU-Pro 16.40 and RX480
All scenes from the cycles benchmark pack, using GPU version of the files with tile size of 2048x2048. Patch applied on top of 818af9c3315cb883436a3d75d634f449133cd3d9. Times are with scene prepartation time removed (so pure rendering time)

Nov 13 2016, 10:29 AM

Nov 12 2016

john peterson (bliblubli) abandoned D2254: 20% to 140% Speedup for the OpenCL kernel.

Splitted the patch in 3 parts with D2339, D2340 and D2341.

Nov 12 2016, 5:03 PM

Nov 11 2016

john peterson (bliblubli) added a comment to T49991: reloading librairies doesn't update node groups.

@Bastien Montagne (mont29) but only on windows, Linux just doesn't really reload/update the node group.

Nov 11 2016, 8:42 PM · BF Blender
john peterson (bliblubli) added a comment to T49991: reloading librairies doesn't update node groups.

yes still happen:

Nov 11 2016, 8:40 PM · BF Blender
john peterson (bliblubli) added a comment to T49991: reloading librairies doesn't update node groups.

Reloading the whole Blend file works as expected.

Nov 11 2016, 12:55 PM · BF Blender
john peterson (bliblubli) assigned T49991: reloading librairies doesn't update node groups to Bastien Montagne (mont29).
Nov 11 2016, 12:54 PM · BF Blender
john peterson (bliblubli) created T49991: reloading librairies doesn't update node groups.
Nov 11 2016, 12:53 PM · BF Blender

Nov 6 2016

john peterson (bliblubli) added a comment to T49952: Cycles- MIS gives better result with resolution superior to input HDRi's resolution.

Thanks for the quick fix Lukas. The image is much cleaner indeed :) And requiring twice the resolution only is much better.

Nov 6 2016, 9:39 PM · Cycles, BF Blender
john peterson (bliblubli) added a dependency for D2341: split-kernel: add new node features to speedup heavy production files.: D2340: Split-kernel: reorganize nodes to get maximal performance on Group 0 (clay render) and better perfs on group 1 and 2..
Nov 6 2016, 6:48 PM
john peterson (bliblubli) added a dependent revision for D2340: Split-kernel: reorganize nodes to get maximal performance on Group 0 (clay render) and better perfs on group 1 and 2.: D2341: split-kernel: add new node features to speedup heavy production files..
Nov 6 2016, 6:48 PM · Cycles
john peterson (bliblubli) created D2341: split-kernel: add new node features to speedup heavy production files..
Nov 6 2016, 6:39 PM
john peterson (bliblubli) added a dependency for D2340: Split-kernel: reorganize nodes to get maximal performance on Group 0 (clay render) and better perfs on group 1 and 2.: D2339: First accepted part of D2254.
Nov 6 2016, 6:14 PM · Cycles
john peterson (bliblubli) added a dependent revision for D2339: First accepted part of D2254: D2340: Split-kernel: reorganize nodes to get maximal performance on Group 0 (clay render) and better perfs on group 1 and 2..
Nov 6 2016, 6:14 PM
john peterson (bliblubli) updated the summary for D2340: Split-kernel: reorganize nodes to get maximal performance on Group 0 (clay render) and better perfs on group 1 and 2..
Nov 6 2016, 6:13 PM · Cycles
john peterson (bliblubli) created D2340: Split-kernel: reorganize nodes to get maximal performance on Group 0 (clay render) and better perfs on group 1 and 2..
Nov 6 2016, 6:08 PM · Cycles
john peterson (bliblubli) updated the summary for D2339: First accepted part of D2254.
Nov 6 2016, 5:22 PM
john peterson (bliblubli) created D2339: First accepted part of D2254.
Nov 6 2016, 5:22 PM
john peterson (bliblubli) edited the description of T49952: Cycles- MIS gives better result with resolution superior to input HDRi's resolution.
Nov 6 2016, 4:55 PM · Cycles, BF Blender
john peterson (bliblubli) added a comment to T49952: Cycles- MIS gives better result with resolution superior to input HDRi's resolution.

True :D
bug mis hdr.blend

Nov 6 2016, 4:54 PM · Cycles, BF Blender
john peterson (bliblubli) updated subscribers of T49952: Cycles- MIS gives better result with resolution superior to input HDRi's resolution.
Nov 6 2016, 4:33 PM · Cycles, BF Blender
john peterson (bliblubli) created T49952: Cycles- MIS gives better result with resolution superior to input HDRi's resolution.
Nov 6 2016, 4:32 PM · Cycles, BF Blender

Nov 5 2016

john peterson (bliblubli) updated the diff for D1598: ImBuf: Add support for WebP image format.

added missing bits and updated to master.

Nov 5 2016, 7:08 PM
john peterson (bliblubli) commandeered D1598: ImBuf: Add support for WebP image format.

updating the patch to correct above mentioned problems and make it to apply on master.

Nov 5 2016, 7:06 PM
john peterson (bliblubli) awarded D2332: Cycles - Distance Culling a Love token.
Nov 5 2016, 6:06 PM
john peterson (bliblubli) triaged T49943: OpenCl rendering error. as "Incomplete" priority.

post the error message here and attach a file to reproduce the problem.
Which driver version are you using? Note that AMD GPU-Pro 16.3 is more stable than 16.4.

Nov 5 2016, 11:33 AM · Cycles, BF Blender

Nov 4 2016

john peterson (bliblubli) closed T49892: Rendering of boolean fail when linked as "Archived".

closing as todo after IRC discussion. Will investigate further what led to this report in cycles.

Nov 4 2016, 4:07 PM · BF Blender
john peterson (bliblubli) added a comment to T49892: Rendering of boolean fail when linked.

updated with renders and viewport images:

Nov 4 2016, 3:50 PM · BF Blender
john peterson (bliblubli) added a comment to T49892: Rendering of boolean fail when linked.

Note that the exporters also do it right, which explains why external renderers render correctly also.

Nov 4 2016, 3:44 PM · BF Blender
john peterson (bliblubli) updated subscribers of T49892: Rendering of boolean fail when linked.
Nov 4 2016, 3:40 PM · BF Blender
john peterson (bliblubli) reopened T49892: Rendering of boolean fail when linked as "Open".

In which case would a dependency not need to be linked and applied like in the original file?

Nov 4 2016, 3:39 PM · BF Blender
john peterson (bliblubli) assigned T49892: Rendering of boolean fail when linked to Bastien Montagne (mont29).
Nov 4 2016, 12:04 PM · BF Blender

Nov 1 2016

john peterson (bliblubli) added a comment to T49613: Single channel byte image takes same amount of memory in Cycles as multi channel image (GPU and CPU).

I would rather close it as todo. There is absolutely no warning in the UI that packed texture will explode the memory usage, even if unpacked afterwards and that an external tool is then needed. Is it really hard to implement single channel packed images that it has to wait for 2.8?

Nov 1 2016, 3:09 PM · BF Blender

Oct 31 2016

john peterson (bliblubli) created T49892: Rendering of boolean fail when linked.
Oct 31 2016, 4:34 PM · BF Blender

Oct 29 2016

john peterson (bliblubli) added a comment to D1864: Improved snapping system.

any news on this patch?

Oct 29 2016, 7:57 PM

Oct 26 2016

john peterson (bliblubli) added a comment to D2254: 20% to 140% Speedup for the OpenCL kernel.

@blibblubli

I might be able to if you agree not to share or distribute or use my setups for commercial use, as all these terrains are to be sold as Asset packs with procedural terrain shaders.

I droped you a PM on Blender artist, If we can work out some protection for me ill give you a link to the file.

Also as im new to Blender code, Im having a nightmare try to apply patches to Blender code on windows, Ive tread command line patch p1 and p2, also tried tortoise svn which also doesn't work. How the hell do I patch code Lol

Oct 26 2016, 10:29 PM

Oct 24 2016

john peterson (bliblubli) added a comment to D2254: 20% to 140% Speedup for the OpenCL kernel.

Sorry, Also this is not micro displace. Experimental turned off.

Multi res mod, Displace mod, modifiers not applied. 150 Samples

Oct 24 2016, 8:39 AM

Oct 23 2016

john peterson (bliblubli) added a comment to D2254: 20% to 140% Speedup for the OpenCL kernel.

@deom damien (dams) The performance boost in production shots definitely comes from the new node_features because all the production shots use the group level 3. I personally didn't notice any impact on performance when just moving cases around. The new organization of the groups is to allow optimal performances in this groups while trying to follow artists usages:

Oct 23 2016, 7:53 PM
john peterson (bliblubli) updated the diff for D2254: 20% to 140% Speedup for the OpenCL kernel.

switched SoA to supported kernel and AoS to experimental. AoS is faster in most cases and benchmark files are set to experimental by default. Point is to gather better time results.

Oct 23 2016, 2:34 PM
john peterson (bliblubli) added a comment to D2254: 20% to 140% Speedup for the OpenCL kernel.

Hi,

There is a simple explanation for this performance boost.

The big switch / case in svm.h is very inefficient and should be replaced by an array of functions.

As this part of the code is called millions of times, millions of comparisons can be avoided.

bliblu bli (bliblubli) has mitigated this problem by placing the most used nodes at the top of the list.

Oct 23 2016, 2:29 PM
john peterson (bliblubli) added a comment to D2254: 20% to 140% Speedup for the OpenCL kernel.

2 good news :) many users on BA and I were not dreaming and AMD fixed the problem in it's latest compiler. Now it would be good if they could ship this new compiler for all platform. Could we have access to some dev drivers through Mai? Or ask during the Bconf as a dev from AMD will be here. It would be good if people working on the OpenCL kernel could have a direct email adress to stop guessing for what is on AMD's side.
Anyway, this patch gives good performance improvements on all cards and platform. Around the 5-15% in my case if we remove the compiler bug.

Oct 23 2016, 10:06 AM

Oct 19 2016

john peterson (bliblubli) added a comment to D2310: Cycles: OpenCL split kernel volume.

It works :) The buggy render seems to be due to the flame attribute having another value mapping than on CPU. Maybe it's not due to this patch?

Oct 19 2016, 8:57 AM

Oct 18 2016

john peterson (bliblubli) added a comment to D2264: Cycles: Support multithreaded compilation of kernels.

Just to say it's a huge time saver for programmers. It allows to test optimisations and new features much faster. So it would be really helpful to add it at least as an experimental feature to master.

Oct 18 2016, 8:19 PM

Oct 17 2016

john peterson (bliblubli) added a comment to D2299: Cycles: OpenCL 3d textures.

Edited: misunderstood patch.

Oct 17 2016, 2:02 PM

Oct 12 2016

john peterson (bliblubli) added a comment to T49593: dxf importer fails to import correctly.

Problem solved, thank you :)

Oct 12 2016, 6:52 PM · Import/Export, Addons
john peterson (bliblubli) added a comment to D2254: 20% to 140% Speedup for the OpenCL kernel.

I tested with a Radeon RX 480 today, on Windows 10 with Radeon driver version 16.9.2 (same as you used). Using the attached microd.blend file:

  • Your blenderartists build: 2m39s
  • My build without patch: 2m28s
  • My build with patch: 2m19s

    I currently don't have good ideas for why you saw such a big speedup, while it's much smaller here with a very similar configuration.
Oct 12 2016, 6:18 PM

Oct 11 2016

john peterson (bliblubli) updated subscribers of T49630: regression MIS Split kernel OpenCL.
Oct 11 2016, 5:20 PM · Cycles, BF Blender
john peterson (bliblubli) created T49630: regression MIS Split kernel OpenCL.
Oct 11 2016, 5:19 PM · Cycles, BF Blender

Oct 9 2016

john peterson (bliblubli) added a comment to D2254: 20% to 140% Speedup for the OpenCL kernel.

I used this script test_single_nodes.py with my patch to get render times with different nodes. Because combinations also have an impact, it makes hard to test every possible combinations (with 84 nodes, we would have to test 2^84 combinations). So I just tested adding single nodes to see there impact when rendering the default cube at 1024spp.
Workflow is to start patched D2254 blender, choose cycles as render engine, set samples to 1, run the script once to compile all kernels. Then render a second time with 1024spp.
It is best to do it with D2264 applied to speedup kernel compilation.
Problem at the moment is that the best method to get pure render time is to get them from the console output, which isn't really easy. So if someone has a solution to expose render time to the python API through an operator or a custom property or whatever, it would really help analyze the times.

Oct 9 2016, 5:24 PM

Oct 7 2016

john peterson (bliblubli) added a comment to T49593: dxf importer fails to import correctly.

@Lukas Treyer (cnd) would be good to add this fix to 2.78a :)

Oct 7 2016, 7:40 PM · Import/Export, Addons
john peterson (bliblubli) added a comment to T49593: dxf importer fails to import correctly.

Thanks :) Fix works for the one error. Note that there is no such big object visible in the original file nor in the DXF when opened with our CAD program, it looks like one point was moved to 0,0,0. The original file being in Germany thus thousands of kilometers away from 0,0,0, it may be what makes the object so big.
I have an narrowed down file with the second error triggering a crash in our production file. bug 3.dxf.zip

Oct 7 2016, 7:36 PM · Import/Export, Addons
john peterson (bliblubli) added a comment to T49583: pdf import fails in 2.78 with latest contrib snapshot.

@Howard Trickey (howardt) thanks a lot for maintaining this very helpful addon :)

Oct 7 2016, 6:39 AM · Import/Export, Addons
john peterson (bliblubli) created T49593: dxf importer fails to import correctly.
Oct 7 2016, 6:38 AM · Import/Export, Addons

Oct 5 2016

john peterson (bliblubli) added a comment to D2254: 20% to 140% Speedup for the OpenCL kernel.

Is it register spill or an if-else hell? Did anyone try to do binary search?

Think question was raised before, binary search is something like:

if(node < NODE_MATH) {
  if(node < NODE_MIX) {
    /* handle nodes from NODE_CLOSURE_BSDF .to NODE_MIX */
  }
  else {
    /* handle nodes from NODE_MIX to NODE_MATH */
  }
}
else {
  /* similar tricks */
}

When there's handful of nodes remained after bisecting, we can roll back to switch() statement. Would be an interesting experiment anyway.

Oct 5 2016, 9:50 PM
john peterson (bliblubli) created T49583: pdf import fails in 2.78 with latest contrib snapshot.
Oct 5 2016, 9:42 PM · Import/Export, Addons

Oct 4 2016

john peterson (bliblubli) added a comment to D2254: 20% to 140% Speedup for the OpenCL kernel.

Little update. Hawaii architectures based cards also greatly benefit from this patch with impressive speedups: See https://blenderartists.org/forum/showthread.php?400121-AMD-RX-480-with-8-Gigs-of-Vram-at-229&p=3100843&viewfull=1#post3100843 .
I did some test to see which nodes trigger a big slowdown on RX480. On the microd.blend at 64spp with one tile I get this rendering time (without scene preparation):

Oct 4 2016, 7:41 PM

Oct 2 2016

john peterson (bliblubli) added a comment to T49508: missing SDL2.dll after installation.

For me it only fails when installing the first time over 2.77a. Repairing or desinstalling then reinstalling resolves the problem.

Oct 2 2016, 7:58 PM · Platform: Windows, BF Blender

Sep 30 2016

john peterson (bliblubli) added a comment to D2239: Signifcant Refactor for the Images as Planes Addon.

Good patch with really useful additions.
It would be good to have an option to set the image plane's origin at the bottom of the image. For trees, people, etc... it's useful to put them on the ground and be able to scale them without them going underground.
An option to add them to a group of the same name as the image/mesh would also allow quick instancing through dupligroups.
I hope this patch gets reviewed and committed quickly :)

Sep 30 2016, 12:19 AM