bug: much area() induces lag even if transform is never changed

[dragoncoder047]

kaplay({
    tagsAsComponents: true,
});

loadBean();

function recur(obj, level) {
    if (level < 0) return;
    const obj2 = obj.add([
        pos(rand(vec2(0), vec2(100))),
        scale(rand(vec2(2))),
        rotate(rand(0, 360)),
    ]);
    if (level === 0) {
        obj2.use(sprite("bean"));
        // obj2.use(area());
    }
    recur(obj2, level - 1);
}

for (var i = 0; i < 500; i++)
    recur(getTreeRoot(), 2);
var frames = 0;
onUpdate(() => {
    frames++;
    debug.log((frames / time()).toFixed());
});

This creates 500 objects with random transforms, with 3 objects nested in each, and on the lowest level, it just adds area(). (commented out above cause less laggy). The transforms are not changing, yet if area() is used it lags much.

Results from kaplyground

Browser	FPS w/o area()	FPS with area()
Chrome 131	120	2 (but it claims 6)
Firefox 132	119	1 (but it claims 5)
Safari 18.1.1	20	0.5 (but it claims 4)

[dragoncoder047]

The strange thing is that this only occurs in the kaplayground, I can't reproduce it when I copy it into a local example in my fork of kaplay, they all get max fps. Even the "export to html" opened via a file:// URL gets lag, but the localhost:8000 version from vs code does not lag.

[dragoncoder047]

I can't reproduce it when I copy it into a local example in my fork of kaplay

Turns out this is because the "new tag methods" broke something , if you change tagsAsComponents: true it lags like expected

[mflerackers]

transforms and areas are going to be cached in the future, so at least they won't be recalculated each frame

[dragoncoder047]

ah good, my game am getting like 5 FPS in Chrome which is the best...

do you have an eta on the caching?

[mflerackers]

I had a branch which was nearly finished, but by now I think there are too many conflicts. So I'll start over.

…

On Mon, 25 Nov 2024 at 08:40, dragoncoder047 ***@***.***> wrote: ah good, my game am getting like 5 FPS in Chrome which is the best... do you have an eta on the caching? — Reply to this email directly, view it on GitHub <#538 (comment)>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/AAHIYWZNSRYUGQW7BE252D32CJPWJAVCNFSM6AAAAABSMMSCHKVHI2DSMVQWIX3LMV43OSLTON2WKQ3PNVWWK3TUHMZDIOJWGMZTOOBWHE> . You are receiving this because you commented.Message ID: ***@***.***>

-- Kind Regards, Marc Flerackers Software Engineer