Page MenuHomeWildfire Games

OptimusShepard (Pirmin Stanglmeier)
User

Projects

User Details

User Since
Feb 24 2017, 5:41 PM (186 w, 5 d)

Recent Activity

Aug 22 2020

OptimusShepard added a comment to D2812: Adds MSAA to anti-aliasing techniques.

Patch works for me (Windows 10, RX 5700).
@vladislavbelov your former patch had AA for trees too, without a big performance hit. Does the new performance impact came from the water bugfix?
Also I don't think, we should care about performance hits by trees, as we have FXAA for lower hardware. I see no need for the implementation in the current state, as FXAA + sharpening looks better than MSAA, as long as trees have aliasing. Especially as FXAA and sharpening has no measurable performance impact compared to MSAA.

Aug 22 2020, 4:25 PM

Aug 8 2020

OptimusShepard added a comment to D2936: Allow limiting the max number of corpses simultaneously visible in the game.
In D2936#128281, @Stan wrote:

Dying animation is the corpse ^^

Aug 8 2020, 6:30 PM
OptimusShepard added a comment to D2936: Allow limiting the max number of corpses simultaneously visible in the game.

A dying animation is not possible here? Or at least a simple visual plop effect?

Aug 8 2020, 6:06 PM

Aug 7 2020

OptimusShepard added inline comments to D2642: Contrast-Adaptiv-Sharpening pass.
Aug 7 2020, 11:30 PM
OptimusShepard awarded D2938: [rc] GL_ARB instancing to reduce draw calls a Like token.
Aug 7 2020, 4:19 PM
OptimusShepard added a comment to D2938: [rc] GL_ARB instancing to reduce draw calls.

gpuskinning is an option in default.cfg that you have to activate using the config file. You could try (de)/activating that

Setting it disabled doesn't change anything.

What are you reporting as being 10/20% slower exactly?

The whole frame.

Edit: I feel like your GC might be too good. Can you try capping max_matrix_uniform to maybe 64 in hwdetect.js ?

That's it. Thx


Render: red non patch, green patched.

Aug 7 2020, 3:53 PM
OptimusShepard added a comment to D2938: [rc] GL_ARB instancing to reduce draw calls.

It should be a straight improvement on svn, so this seems odd. How did you replay exactly? Were you using GPU skinning? Were you using prefer glsl? What map?

I always use a 200 units replay on sahydian buttes, camera will not moved or zoomed. What is GPU skinning? GSLS enabled.

Edit -> Also what is your graphics card?

Radeon RX 5700

Aug 7 2020, 3:15 PM
OptimusShepard added a comment to D2938: [rc] GL_ARB instancing to reduce draw calls.

I also got this compiler warnings
\source\renderer/ModelVertexRenderer.h(158): warning C4100: "model": Unreferenzierter formaler Parameter (Quelldatei wird kompiliert ..\..\..\source\renderer\InstancingModelRenderer.cpp)
\source\renderer/ModelVertexRenderer.h(158): warning C4100: "shader": Unreferenzierter formaler Parameter (Quelldatei wird kompiliert ..\..\..\source\renderer\InstancingModelRenderer.cpp)
\source\renderer/ModelVertexRenderer.h(158): warning C4100: "model": Unreferenzierter formaler Parameter (Quelldatei wird kompiliert ..\..\..\source\renderer\HWLightingModelRenderer.cpp)
\source\renderer/ModelVertexRenderer.h(158): warning C4100: "shader": Unreferenzierter formaler Parameter (Quelldatei wird kompiliert ..\..\..\source\renderer\HWLightingModelRenderer.cpp)
\source\renderer/ModelVertexRenderer.h(158): warning C4100: "model": Unreferenzierter formaler Parameter (Quelldatei wird kompiliert ..\..\..\source\renderer\ModelRenderer.cpp)
\source\renderer/ModelVertexRenderer.h(158): warning C4100: "shader": Unreferenzierter formaler Parameter (Quelldatei wird kompiliert ..\..\..\source\renderer\ModelRenderer.cpp)

Aug 7 2020, 3:02 PM
OptimusShepard added a comment to D2938: [rc] GL_ARB instancing to reduce draw calls.

I made some quick profiling. I noticed that the amount of draw calls is halfed. I also tested value 1 against 64. 64 is much faster. But compared to the non patched replay, there was a performance hit by ~10-20%.
Does this hit come from your new profiler option "saved draw calls"? I think some calculations of profiler are active, even if you don't have open the profiler panel. Do you have a non profiler version to test?

Aug 7 2020, 2:52 PM

Aug 6 2020

OptimusShepard added inline comments to D2936: Allow limiting the max number of corpses simultaneously visible in the game.
Aug 6 2020, 3:09 PM
OptimusShepard added inline comments to D2936: Allow limiting the max number of corpses simultaneously visible in the game.
Aug 6 2020, 2:53 PM
OptimusShepard awarded D2936: Allow limiting the max number of corpses simultaneously visible in the game a Like token.
Aug 6 2020, 2:52 PM

Jul 25 2020

OptimusShepard added inline comments to D2642: Contrast-Adaptiv-Sharpening pass.
Jul 25 2020, 12:08 AM
OptimusShepard added inline comments to D2642: Contrast-Adaptiv-Sharpening pass.
Jul 25 2020, 12:06 AM

Jul 24 2020

OptimusShepard updated the diff for D2642: Contrast-Adaptiv-Sharpening pass.

Edited the sharpness initialization. Disables sharpening, if post progressing, or GLSL, is disabled.

Jul 24 2020, 11:55 PM

Jul 23 2020

OptimusShepard added a comment to D14: [RC] Thread the pathfinder computations.
In D14#125536, @wraitii wrote:

The smallest "unit" of computation is a single path, so to get an improvement with 16 threads you'd need more than 16 paths per turn. This is possible in intense combat situation (e.g. combat demo huge), but not very likely.

Jul 23 2020, 9:14 AM
OptimusShepard added a comment to D14: [RC] Thread the pathfinder computations.
In D14#125537, @Stan wrote:

He has 8 cores 8threads :)

Jul 23 2020, 9:06 AM

Jul 22 2020

OptimusShepard added inline comments to D2642: Contrast-Adaptiv-Sharpening pass.
Jul 22 2020, 11:14 PM
OptimusShepard updated the diff for D2642: Contrast-Adaptiv-Sharpening pass.

Updated the versioning of the used GLSL version.

Jul 22 2020, 11:09 PM
OptimusShepard added a comment to D14: [RC] Thread the pathfinder computations.

I made some profiling with different numbers (1, 2, 3, 4, 5, 6, 8, 12, 16) of threads. If I compare the results, I guess, more than 16 threads wont improve the performance further.

Jul 22 2020, 7:59 PM

Jul 18 2020

OptimusShepard added a comment to D2857: Matrix3D SSE.
In D2857#123243, @Stan wrote:

Can you figure out what's causing the spikes?

Didn't find the cause yet. But I recognized, that the profiler gains these pikes very much. Without the framedrops were much lower.

Jul 18 2020, 12:48 AM

Jul 10 2020

OptimusShepard added a comment to D2857: Matrix3D SSE.
In D2857#123243, @Stan wrote:

I like your approach better but interestingly it's done slightly differently in https://github.com/0ad/0ad/blob/d3e68a99e7f715ad7921a81e959f8ac51dfa1248/source/graphics/Color.cpp

Oh, I think they changing the instructions by runtime, doesn't they? A bit ugly, I think, but we could use FMA :)

Can you figure out what's causing the spikes?

I will try, but on an first look, the profiler doesn't show me something useful.

Jul 10 2020, 2:18 PM
OptimusShepard updated the diff for D2857: Matrix3D SSE.

Updated the year of the license header.

Jul 10 2020, 10:34 AM

Jul 9 2020

OptimusShepard updated the diff for D2857: Matrix3D SSE.

Including the SSE header, because Vulcan fails for the none Windows tests.

Jul 9 2020, 11:55 PM
OptimusShepard updated the diff for D2857: Matrix3D SSE.

Removed the AVX and FMA version, as we don't be able, to change instructions by runtime. Furthermore the AVX instructions aren't faster than SSE here.

Jul 9 2020, 11:32 PM
OptimusShepard added a comment to D2857: Matrix3D SSE.

I have rewrite the patch, so it uses only SSE. That I have used for the profiling. I will upload it later this day.

Jul 9 2020, 3:40 PM
OptimusShepard added a comment to D2857: Matrix3D SSE.

I tested the build flags, SSE seems to be the only flag with an positiv impact. AVX2 makes everything worse. I also made some profiling.
Current version:

Jul 9 2020, 3:38 PM

Jul 3 2020

OptimusShepard planned changes to D2857: Matrix3D SSE.

The hardware request doesn't work, like it should.

Jul 3 2020, 1:04 AM
OptimusShepard created D2857: Matrix3D SSE.
Jul 3 2020, 1:02 AM

Jun 18 2020

OptimusShepard added a comment to D2780: Bugfix: GLSL query for FXAA .

Also isn't PostProc locked by GLSL? Since we don't have effects for ARB (iirc).

No it isn't. It's currently independent. I have also tried to lock postproc, to lock antialiasing, but that doesn't work. It would only lock postproc. A bit strange I think.

Jun 18 2020, 9:38 PM
OptimusShepard updated the diff for D2780: Bugfix: GLSL query for FXAA .

Edit some comments.

Jun 18 2020, 9:31 PM

Jun 14 2020

Krinkle awarded D2789: SSE Vector3D a Orange Medal token.
Jun 14 2020, 5:40 PM

Jun 13 2020

OptimusShepard added a comment to D2642: Contrast-Adaptiv-Sharpening pass.

Shouldn't the CAS license be mentioned in LICENSE.txt?

Jun 13 2020, 1:50 PM
OptimusShepard updated the diff for D2642: Contrast-Adaptiv-Sharpening pass.

Add license, change header year.

Jun 13 2020, 1:49 PM
OptimusShepard updated the diff for D2789: SSE Vector3D.

Add context.

Jun 13 2020, 1:16 PM
OptimusShepard updated the diff for D2789: SSE Vector3D.

The current version is not profiled yet, so I do not know, which of these functions will improve performance.
"-" and "+" seems to have grate impact to the game, when many units are in the field of view. Until yet I didn't get it profiled, when many units are used.

Jun 13 2020, 11:57 AM

Jun 5 2020

OptimusShepard changed the visibility for D2789: SSE Vector3D.
Jun 5 2020, 10:19 PM
OptimusShepard changed the visibility for D2789: SSE Vector3D.
Jun 5 2020, 10:00 PM
OptimusShepard created D2789: SSE Vector3D.
Jun 5 2020, 9:53 PM

Jun 1 2020

OptimusShepard created D2780: Bugfix: GLSL query for FXAA .
Jun 1 2020, 9:23 PM

May 31 2020

OptimusShepard updated the diff for D2642: Contrast-Adaptiv-Sharpening pass.

Updating the licence and the link, as AMD had updated it. Changing the game options name to the AMD recommented one.

May 31 2020, 10:21 AM

May 20 2020

OptimusShepard added a comment to D2726: AMD Ryzen fix.

I have tested with different setups.

May 20 2020, 6:40 PM

May 3 2020

OptimusShepard updated the diff for D2642: Contrast-Adaptiv-Sharpening pass.

Modified it, so it could be used on all maps.

May 3 2020, 12:59 PM
OptimusShepard accepted rP23484: Adds anti-aliasing option with FXAA algorithm..
May 3 2020, 12:11 PM

Apr 28 2020

OptimusShepard added inline comments to D2642: Contrast-Adaptiv-Sharpening pass.
Apr 28 2020, 10:54 AM
OptimusShepard added inline comments to D2642: Contrast-Adaptiv-Sharpening pass.
Apr 28 2020, 10:05 AM
OptimusShepard updated the diff for D2642: Contrast-Adaptiv-Sharpening pass.

Revert to previous version and edit some comments.

Apr 28 2020, 9:56 AM

Apr 27 2020

OptimusShepard added a comment to D2642: Contrast-Adaptiv-Sharpening pass.

First time using Arcanist. Something went wrong, files are missing. How can I turn back to previous version?

Apr 27 2020, 11:41 PM
OptimusShepard updated the diff for D2642: Contrast-Adaptiv-Sharpening pass.
  1. Updating D2642: Contrast-Adaptiv-Sharpening pass #
  2. Enter a brief description of the changes included in this update.
  3. The first line is used as subject, next lines as comment. #
  4. If you intended to create a new revision, use:
  5. $ arc diff --create
Apr 27 2020, 11:23 PM
OptimusShepard added inline comments to D2642: Contrast-Adaptiv-Sharpening pass.
Apr 27 2020, 2:12 PM
OptimusShepard planned changes to D2642: Contrast-Adaptiv-Sharpening pass.

Waiting until D2712 is committed. A little modification of the code of this patch has to be done.

Apr 27 2020, 10:05 AM
OptimusShepard accepted D2712: Fixes FXAA for maps with default effects.

Patch is working as expected. All tests passed as it should. Code looks correct.
Accepted. Thank you Vladislav

Apr 27 2020, 10:00 AM

Apr 20 2020

OptimusShepard added a comment to D2642: Contrast-Adaptiv-Sharpening pass.

The patch is waiting for reviewing, what update do you want to have? The patch is mostly working. As it have the same problem as the FXAA patch, that it doesn’t work on all maps, and I don't know how to solve it, I will wait and see, how Vladislav is solving this problem.
If you want to test it, you could choose e.g. the new shirmish maps Obedska Bog or Sahyadri Buttes.

Apr 20 2020, 10:08 AM

Apr 13 2020

OptimusShepard added a comment to D2699: Fixes FXAA on some platforms.

Works on Intel UHD 620 with Windows.

Apr 13 2020, 9:31 PM

Apr 12 2020

OptimusShepard added a comment to D2699: Fixes FXAA on some platforms.

Patch is working on my AMD Radeon RX 570 Manjaro Linux and Windows system.

Apr 12 2020, 10:55 PM

Mar 27 2020

OptimusShepard added a comment to rP23484: Adds anti-aliasing option with FXAA algorithm..
In rP23484#41603, @Stan wrote:

FXAA doesn't work on my linux system. On dual boot Windows it's working.
Manjaro 64Bit, Radeon RX 570

Same map? It doesn't work on all maps if postprocessing is disabled on the map settings...

Mar 27 2020, 8:29 PM
OptimusShepard raised a concern with rP23484: Adds anti-aliasing option with FXAA algorithm..

FXAA doesn't work on my linux system. On dual boot Windows it's working.
Manjaro 64Bit, Radeon RX 570

Mar 27 2020, 8:21 PM

Feb 26 2020

OptimusShepard added a comment to D2642: Contrast-Adaptiv-Sharpening pass.

Thx

Feb 26 2020, 6:52 PM
OptimusShepard updated the diff for D2642: Contrast-Adaptiv-Sharpening pass.

Change the representation of floats in cas.fs.

Feb 26 2020, 6:51 PM
OptimusShepard added inline comments to D2642: Contrast-Adaptiv-Sharpening pass.
Feb 26 2020, 3:59 PM
OptimusShepard updated the diff for D2642: Contrast-Adaptiv-Sharpening pass.

Add an editing info to the cas.fs header.

Feb 26 2020, 3:51 PM

Feb 25 2020

OptimusShepard added inline comments to D2642: Contrast-Adaptiv-Sharpening pass.
Feb 25 2020, 8:51 PM
OptimusShepard added inline comments to D2642: Contrast-Adaptiv-Sharpening pass.
Feb 25 2020, 8:44 PM
OptimusShepard updated the diff for D2642: Contrast-Adaptiv-Sharpening pass.

Add slider for sharpness factor.

Feb 25 2020, 8:25 PM

Feb 24 2020

OptimusShepard added a comment to D2642: Contrast-Adaptiv-Sharpening pass.

I use Intel CPU, and the graphics card USES NVIDIA GTX850M. The operating system is Windows10 64bit. Do you support Intel CPU and NVIDIA series of graphics CARDS in this patch

Feb 24 2020, 8:25 AM

Feb 23 2020

OptimusShepard updated the diff for D2642: Contrast-Adaptiv-Sharpening pass.

Replace rcp(x) to 1/x.
rcp seems to be a performance optimized function for AMD GPUs. As strange, rsqrt is also a special function, which doesn't work with my GPU. I had it replaced bevor my first upload.
@vladislavbelov do you know more about such things? Is it possible to implement a parallel path for AMD hardware?

Feb 23 2020, 4:57 PM
OptimusShepard added a comment to D2642: Contrast-Adaptiv-Sharpening pass.
In D2642#110914, @Stan wrote:

AFAIK @gameboy uses Windows.

Feb 23 2020, 4:19 PM
OptimusShepard retitled D2642: Contrast-Adaptiv-Sharpening pass from Contras-Adaptiv-Sharpening pass to Contrast-Adaptiv-Sharpening pass.
Feb 23 2020, 4:17 PM
OptimusShepard added a comment to D2642: Contrast-Adaptiv-Sharpening pass.

@OptimusShepard I tested your patch and when I select CAS in the options, the error appears:

ERROR: Failed to compile shader 'shaders/glsl/cas.fs': 0(64) : warning C1503: undefined variable "rcp" 0(70) : warning C1503: undefined variable "rcp" 0(73) : warning C1503: undefined variable "rcp" 0(64) : error C1008: undefined variable "rcp" 0(70) : error C1008: undefined variable "rcp" 0(73) : error C1008: undefined variable "rcp"

Feb 23 2020, 4:12 PM

Feb 21 2020

nani awarded D2642: Contrast-Adaptiv-Sharpening pass a Like token.
Feb 21 2020, 12:02 AM

Feb 20 2020

OptimusShepard planned changes to D2642: Contrast-Adaptiv-Sharpening pass.

I will add a slider for selecting the sharpness factor.

Feb 20 2020, 9:00 PM
OptimusShepard added a comment to D2642: Contrast-Adaptiv-Sharpening pass.

Do I have to add something to the LICENSE.txt?
I ported the reshade port of AMDs CAS, so is there anything else to consider?

Feb 20 2020, 8:55 PM
OptimusShepard created D2642: Contrast-Adaptiv-Sharpening pass.
Feb 20 2020, 8:50 PM

Feb 19 2020

OptimusShepard abandoned D1253: Better ultrawidescreen support in game.

The ultra widescreen support should be taken into account in a future revision of the gui.

Feb 19 2020, 10:08 PM

Feb 15 2020

OptimusShepard added a comment to D2528: Performance improvements to VertexBuffer.

The game is still running correctly. I recognize an performance improvement up to 10%.

Feb 15 2020, 6:48 PM
OptimusShepard added a comment to D2621: Adds GLSL shaders of overlay lines.

The game does run without problems. I can't recognize any difference between with/without "Prefer GLSL".

Feb 15 2020, 6:41 PM

Jan 26 2020

OptimusShepard added a comment to D2537: Adds anti-aliasing option.

As I know AMD has made his new Radeon Image Sharpening filter free open source, I would recommend to have a look on it.

Jan 26 2020, 2:37 PM

Jan 9 2020

OptimusShepard added a comment to D2537: Adds anti-aliasing option.

I would suggest to split the anti-aliasing and the sharpening filter in two seperat options in the settings. Also I like the idea to have a slider that everyone could take is favor settings. As I know AMD has made his new Radeon Image Sharpening filter free open source, I would recommend to have a look on it. The quality is very impressive.

Jan 9 2020, 1:07 PM

Dec 8 2019

OptimusShepard added a comment to D2353: L3 Ryzen 3000 Bugfix.

I've made a mistake by testing procedure. So I didn't seen that the OOM only occurs when I used a version which I compiled myself. So the bug is related to my Visual Studio settings or something like that. Sorry for that.

Dec 8 2019, 7:00 PM
OptimusShepard updated the diff for D2353: L3 Ryzen 3000 Bugfix.

Added comment with link.

Dec 8 2019, 2:48 PM

Dec 4 2019

OptimusShepard added a comment to D2353: L3 Ryzen 3000 Bugfix.

I tested the game with, and without the patch for several times. With the patch the game crashes with an OOM after 30-45min. The OOM report isn't every time the same. Without the patch the game crashes only at the start. I have tested it with following settings. Mapsize big, very huge start ressources, 4 ai vs 4 ai, anatolien plateau, speed 20x.
In my opinion someone should take the current documentation of AMD and make a clean implementation.

Dec 4 2019, 9:25 PM

Dec 3 2019

OptimusShepard added a comment to D2353: L3 Ryzen 3000 Bugfix.

Can you test if the OOM is reproductible though?

Dec 3 2019, 8:12 PM

Nov 30 2019

OptimusShepard added a comment to D2353: L3 Ryzen 3000 Bugfix.

I currently have no time to write the patch, sorry. If someone else writes this patch I would test it.
Here are the required information which Imarok got from the AMD forum.

Nov 30 2019, 12:28 PM

Oct 10 2019

OptimusShepard added a comment to D2353: L3 Ryzen 3000 Bugfix.
In D2353#98805, @Stan wrote:

@OptimusShepard will you update the patch ? :)

Oct 10 2019, 10:26 AM

Oct 5 2019

OptimusShepard added a comment to D2353: L3 Ryzen 3000 Bugfix.
In D2353#98463, @Imarok wrote:
In D2353#98462, @elexis wrote:
In D2353#98374, @Imarok wrote:

I think we now had multiple Zen 2 users complaining about crashes.

Got a link?

Just see the ticket and the link in the post above.

Here is another link. https://wildfiregames.com/forum/index.php?/topic/15796-known-problems-please-read-before-posting/&do=findComment&comment=384917
So we allready have Ryzen 3600, 3600X, 3700X and 3800X. So we only miss a crash report of an 3900X user to complete the Ryzen Zen2 series.

Oct 5 2019, 11:41 AM

Oct 3 2019

OptimusShepard added a comment to D2353: L3 Ryzen 3000 Bugfix.
In D2353#98365, @elexis wrote:

This #4360?

Oct 3 2019, 10:17 PM
Stan awarded D2353: L3 Ryzen 3000 Bugfix a Hungry Hippo token.
Oct 3 2019, 10:12 PM
OptimusShepard created D2353: L3 Ryzen 3000 Bugfix.
Oct 3 2019, 10:04 PM

Mar 22 2019

OptimusShepard added a comment to D1758: Smooth new Kushite background.

Before the patch the balkon position was wrong.


Now after the patch the balkon is on the right position but stretched.

Mar 22 2019, 8:05 PM · Contributors

Jan 7 2019

OptimusShepard accepted rP21378: Allow units to be positioned with freehand placement.
Jan 7 2019, 2:28 PM

Oct 22 2018

OptimusShepard created D1656: Bugfix rP21378.
Oct 22 2018, 6:34 PM

Oct 18 2018

OptimusShepard added a comment to D1652: Freehand formation position .

Oct 18 2018, 6:34 PM
OptimusShepard planned changes to D1652: Freehand formation position .

hotkey "alt" for change to freehand formation is not implemented yet. It's currently activated every time.

Oct 18 2018, 6:32 PM
OptimusShepard created D1652: Freehand formation position .
Oct 18 2018, 6:30 PM

Oct 17 2018

OptimusShepard added inline comments to rP21378: Allow units to be positioned with freehand placement.
Oct 17 2018, 12:03 PM

Oct 14 2018

OptimusShepard raised a concern with rP21378: Allow units to be positioned with freehand placement.
Oct 14 2018, 8:19 PM

Sep 10 2018

OptimusShepard added a comment to D1624: Smoothing Wild Lake.

Good Point. I have quickly tested it for every map size. Seems to be that docks could be built on the whole shoreline.

Sep 10 2018, 6:07 PM
OptimusShepard added a comment to D1624: Smoothing Wild Lake.

I know what your mean, but I think this isn't necessary. "5" looks like a good compromise on all maps I think. Interestingly, this problem doesn't occur on gigantic map size.
Maybe someone will later rework the complete map, there are a little bit too much animals, and there are also some unreachable trees. Until then this patch should work fine I think.

Sep 10 2018, 5:07 PM
OptimusShepard created D1624: Smoothing Wild Lake.
Sep 10 2018, 10:39 AM

Jul 4 2018

OptimusShepard added a comment to D1590: 0 A.D. Empires Ascendant Multiplayer Lobby Privacy Policy.
In D1590#63859, @elexis wrote:

Don't you gather hardware informations? I didn't read something like that.

That's the UserReporter seen in the main menu.
That is not sent to the Multiplayer Lobby service and it's not related to the lobby either.
It is sent in the main menu, has custom terms (we didn't check that for GDPR complience yet), sent for singleplayer matches too
and the sent data does as far as I know not contain the lobby username nor an IP address but a unique userreporter ID saved in the users config file.

So you raise an important point, users should find out how the personal data of other service affects their rights.
The lobby terms should probably link to a document with references to the terms of the different places.

Jul 4 2018, 7:22 PM