If you've read Scott's GPU reviews, you probably have some understanding of the functional units inside Nvidia's latest graphics chips. A recent blog post from the company sheds additional light on how those resources come together to draw pixels on the screen.
The post steps through the logical pipeline, starting with an initial draw call and culminating with the final pixel. It explains how work is distributed between the various units in detail, and it includes some block diagrams I haven't seen before.
While the post doesn't unearth fresh nuggets about Nvidia's core GPU architecture, it's recommended reading for anyone who wants to learn more about how modern graphics chips work. Those who want to dive deeper can follow the post's links to more technical whitepapers on Fermi, Kepler, Maxwell, and other subjects.