The Weakest Spot Of The AMD Gallium3D Driver

In the discussion about the latest AMD R600g driver improvements by Marek Olšák, the prolific independent contributor shares some of his personal views on the open-source graphics driver itself.

There's this forum thread talking about the latest work done by Marek Olšák, which was cleaning up and re-factoring parts of the driver -- just one of many accomplishments by the student developer for the open-source AMD driver.

When asked by a Phoronix reader if there was anything he wish he could do to the R600g code-base and if he had an infinite amount of time to work on any one piece of the driver, like a serious core design flaw, he had an interesting response:

Nothing big comes to my mind right now. R600g certainly needs a good optimizing compiler, it's the weakest spot of the driver. Most of the design flaws have been either fixed already or are in the process of being fixed.

From Marek's perspective, the biggest issue is the lack of a good optimizing compiler for R600g, which would be something big to tackle.

Another Phoronix reader then chimed in, "Jérome said, more than one year ago, that the kernel interface is quite bad and is (or will be) a bottleneck. But it's really hard to heavily modify this." Marek's response to this was, "I think the kernel interface is quite good. Jerome just likes to rewrite things from scratch. It's a sport for him. Most of the kernel code is executed in another thread and runs in parallel with Mesa most of the time. You wouldn't probably even notice if the kernel code were twice as slow."

Join the conversation.