this make totally no sense! anyone can explain to me da why? i simply double the fps without any major change.
>>110061815Are you retarded? In xvfb it obviously rendering at lower resolution
>>110061826> lower resolutionThey are rendered the same window size, i just resize smaller them in video to reduce the cpu use.
>>110061837nigger your xvfb already rendering at lower resolution, and then you moved vkcube to 1/4 of that resolution, and then you're surprised that it's faster than vkcube that's rendering on 1/4 of your native resolution lmao
>why does it render faster if it doesn't have to render most of the sceneo tempora, o mores
>>110061847how about this? triple the fps at larger screen? glitches?
>>110061854What do you means? It is constant and stable, double fps of course
I think i would next test the openGL demo to see whether it is a glitch or it is not.
>>110061815You mean you get higher fps when the cube is hidden?
>>110062267> when the cube is hiddenYes fps does increase a little bit when it is hidden, but not to the point of double the fps. I think there is something i missed here.
Btw the general consensus is that it should be slower since i’m adding more code into the entire pipeline, which add to more work need to be done thus resulting in much lower fps, but it is to my surprise is totally opposite.
opengl test is normal. but not the other one.
>>110061815>this make totally no senseyou resize the window, part of the object not being visible, this will increase fps because there's less things for the cpu to calculate and draw. bigger window, more of the object visible on the screen = more things to calculate = less fps. basics of most renderers. unfortunately you're trying to render 3d objects on a device that's simply too shit for any high frame rates.
>>110062487You check this>>110062345I resize the window in another test too but it doesn’t necessarily get you smaller screen for bigger fps, it is opposite btw.
side by side comparison video, resized to match each other in resolution.
Bump
>>110062844Your fps metrics are wrong.you are measuring the fps of xvfb but the terminal actually tells you the real fps.You can also just inspect the video frame by frame and you will see that it updates at the same time (it's possible the recording is at a fixed fps)Also while testing opengl, you are using a different fps metric which does not have the same flaw.
>>110065506I just look at the number, about the smoothness, yup it’s quite choppy, I think the cpu is unable to catch up with all the work. I think i can try something else.
