|How to best achieve encapsulation on the GPU?||
I wrote a blog about the issues we're seeing writing good maintainable software for the GPU. However, the blogging software on gpucomputing.net makes the entire thing unreadable (lots of code examples). So I'm just going to post a link: iheartcode.blogspot.com/2010/04/encapsulation-on-gpu.html
I'm curious to hear feedback from other GPU developers. Am I missing an obvious trick? Or is this a guiniune fundamental flaw in how we currently program GPUs? I'm hoping to start a discussion about best coding practices for the GPU.