SnipG's Forum Posts

  • Maybe they are 2d fluids, they are made to look like particles. I doubt its full physic engine or shaders

  • This has been happening even before 120/121 etc. In some strange cases webgl just crashes in preview(results black screen) and then editor view is blank, even tought you can still click and select objects. Ordinary it happens when doing some tests with huge amount of sprites.

    You can get similar crash in shadertoy.com when viewing some examples there.

    It has never happened with just 1 preview test, but over time while you have done like 5-6+ test. So there could be some kind of leak.

    I got different console error when it has happened. "WebGL: "Context_Lost_Webgl: loseContext: context lost"

  • C3 runtime uses WebGL2 while C2 runtime uses Webgl1. So your browser does not support Webgl2 or it has been disabled.

  • I reported issue in 118 about LoS it got fixed in 119. It was combination of LoS + Obstacle(tilemap in my case). So try maybe deleting some Obstacles and if you find one which fixes it, you can recreate and report the issue.

  • If you use worker then try without it, if gpu % goes higher then cpu %(with worker) it can crash. You can check it in canarly with gpu % and cpu %.

  • Did some quick tests.

    Sprite with bullet + fade. Before (111 stable)- C2 ~10400 sprites vs C3 5500 sprite at ~30 FPS.

    After update (112 Beta) - C2 ~10 400 vs C3 around same as C2 ~10 000, but seems maybe 2-3% slower.

    Bullet behavior on sprite (111 stable) - C2 ~ 11000 vs C3 ~9500. (112 beta) - C2 ~11000 vs C3 ~ 14 000

    Did some extreme test. Sprite with fade,bullet,Solid,Flash,DestoryOutsideLayout,Rotate:

    (111 stable) C2 ~7700 vs C3 ~ 1800. (112 beta) C2 ~ 7700 vs C3 6700.

    When sprite has 1 behavior it is better in 112 Beta and in 111 in most cases, but if multiple behaviors are added, C3 is same or bit slower in come tests.

    Just Fade on sprite, is around same in 111 stable and 112 beta. C2 - 15 000 vs C3 ~25 000.

    Sine same in 111 and 112: C2 ~ 23 000 sprites vs C3 ~ 34 000.

    Test are results with 30 fps, so when you try achieve 60fps, results maybe be different. Done it with my own tests to give some overview.

    C3 peforming a bit slower is to do with C3 trying to match C2 runtime, so it is faster but doing extra work to match C2.

    It seems really nice that C3 is getting better with every update.

  • Is layout grid bugged or which grid is running out for you ?

  • It is most likely a typo, it should be inside that previous event. If keyboard or gamepad buttons are not down, set walking to false once. Atm trigger once is seperated, it sets walking to false each tick, which is bad behavior.

  • I tried chrome bug report apk on samsung S9, with default browser chrome or samsung internet 7.2.10.33. Can't reproduce it. Does anyone even have some solid example that fails, or some game example?

  • They should be exactly the same, but with different functionality. I tried it out with some bunnymark test that circulated forum sometime ago. It was just plank test with spawn and moving, if you add frames etc then i don't know how much difference it will have.

  • Disable preview effects in editor.

  • I tried out some versions with family and container. Family version on my pc had 1000 obj ran 48-50fps, container ~40, i mean just rotate. After last C3 update it sure does run faster on C3.

    My test could had mistakes, but maybe if container had own variable, could it run same fast as family or it makes no difference?

    Btw thous example projects in this post have disabled solid on base sprite, which doubles you cpu usage vs if you remove it, only if you enable car movement.

    Anyway could someone with Chowdren do some optimized test to?

  • You don't need any super Rig to run it, when object does not move it won't take much CPU. I tested with this: Maybe

    i removed something to much, test yourself. https://drive.google.com/file/d/1Ixjasz ... sp=sharing

    Renderer did not give any fps boost at current proj, but spawning cars at higher obj count was much faster with that. Also new renderer layout rotate was bugged with this example, and did not rotate.

  • Very nice, is cool when someone does some extensive stuff. But you should try it on with some default template or something similar for benchmark. Clearly that default fake3d is very strangely optimized.

    I changed that fake3d and got some results to.

    Default fake3D

    My own: had solid and car enabled.

    Test1 Behaviors enabled

    My own: solid and car disabled.

    Test2 Behaviors disabled

    When rotating that 1000 cars fps drops 15-25

    I did this on Construct 3

  • Try Construct 3

    Develop games in your browser. Powerful, performant & highly capable.

    Try Now Construct 3 users don't see these ads

  • You could achieve that with physic behaviour, apply force and rotate. But do you know what game that is, game name, you have linked here (https://giant.gfycat.com/IndolentQuarterlyAruanas.webm)?