Image Rendering Test

Question

I am benchmarking a custom brower and want to benchmark the rendering speeds of different types of images (gif, jpg, png) of the same file size to see which of the image formats this browser renders the fastest.

My process was just to have a simple seperate HTML page for each type of image and just use a Javascript counter before it is rendered and and after to measure the browser's rendering speed of that specific image.

Any thoughts on this process? Any thoughts on how to improve it?

Answer 1

Well, it's difficult to get meaningful generic results that way. You're measuring a combination of loading html, javascript and an image. Depending on where you're loading them from, you're also measuring the disk or network cache. The image rendering code is going to have some startup time, is dependent on a memory allocator and possibly garbage collection. Then there is image size, color depth, amount of compresion, number of images on the page, scaling, the influence of style sheets, the resolution of the javascript timer. Oh, and are you rendering to a visible part of a window, in a layer, or off-screen.

But don't worry, you'll be able to come up with a usable test. For your specific situation. Or the differences might even be very clear.