so lately ive been working on a few projects with the use of lwjgl, but have ran into some problems with people being able to run my projects(just 2 reported cases)
i cannot be sure that the problem is not with my code but i ran all my programs on about 12 computers and it was worked on 10… so yeah i dont think it is an issue with the client.
Anyway i figured since one of these 2 computers is of a very good friend and its crucial that he help me test my program that i would put in a fall back to Java2d of my program . ive gone and abstracted all my graphics calls and rendering mechanisms but ive come to a dilemma.
In looking for the best ways to render with java2d i found a plethora of methods but not all seem viable. a lot of people suggest i do per pixel rendering onto a buffered image and then g.draw that image because(and this is their logic) making calls to g.drawImage has a lot of overhead.
so what you do with this pixel method is make a new image and draw straight into the raster using a pixel[] array. We do this by looking through each pixel and blah blah blah you get it.
contrary to what i read online, this method was very very slow.
running uncapped rendering one image onto the screen using this method my fps dropped to around 400 fps as opposed to 2300fps when using the g,drawImage.
so currently i am sticking with g.drawImage to draw every single object since its way faster than looping through 40k pixels each frame.
Now my question is . is there some way to optimize rendering images with Java2D? from personal experience or prior knowledge what can you guys come up with?
Thanks in advance