Hello, we are developing a project for HTML5 and need to optimize as much as possible. As a Project artist I have always worked with the limitations in mobile dev.
I wonder if the statistics we have are correct, what do you think ?
Is the difference between the Draw and GPU normal?
when performing outer tests without UE4 editor make sure to check release type of packaged game and grab results from different hardware on your choise (like FPS, GPU load, GPU RAM load), because debug version may slow down game same as running game from editor
process monitor, when you have packaged release type game (for test, not right before actual release) you can add filter to exact process name and after playing game some time open summary file usage on HDD to check if your release game access files on HDD too often, too long, because then it may mean UE4 have some dark corners in optimization itself and you can try make same visual things in other way to check if too heavy file usage changes (first time i saw such was a post about world of tanks, of course it’s not like any html5 game, but his problem was 30 seconds load level, because it was like 0.5 gb and he solved it in hardcore way, he just made vrtual RAM drive and putted that big file here, so when game attempted to read it from HDD, virtual RAM drive actualy gave files from RAM and of course speed difference between RAM, HDD and even SSD still too much, RAM fater then SSD still like 30 times)
hope these tips help if you look for harcore optimization, think not only about how straight GPU load or GPU RAM load can affect optimization, but how side things can affect it too, like different windows version (some people tell windows 10 makes more FPS in games that uses too much CPU), disabling windows aero/full screen mode/borderless window in windowed mode, can never know how bunch of plenty hidden technologies behind things we don’t see work together and affect each other
