Skip to main content
3 events
when toggle format what by license comment
Apr 12, 2011 at 15:38 vote accept Hannesh
Apr 12, 2011 at 12:05 comment added Kylotan Just to expand on this, the historical reason why framebuffer access has been slow is because the instructions are very heavily pipelined. Getting access to a given framebuffer pixel would involve stalling the current pipeline until all the other pipelines were flushed to complete any rendering relevant to the queried pixel. Even in the odd case of a completely non-parallel GPU, you'd still be flushing the entire pipeline for each query, which is just a bad idea. No doubt things are a bit different in the world of programmable hardware now but I expect a similar principle applies.
Apr 11, 2011 at 16:40 history answered user1430 CC BY-SA 3.0