Benchmark performance impact of no_gfx_api vs traditional rendering

[jkelin]

I would love to see how no_gfx_api which is bindless, uses GPU pointers and has global barriers compares to traditional rendering. It would also be nice to show how much code is it saving compared to Vulkan implementation of the same features!

[leotmp]

Sorry for late reply but yeah this would be cool to do eventually. The LOC comparison part could maybe be inserted into the web page docs I'd like to write