Monday, June 16, 2014

State of GL 4.x revealed via "apitest" benchmark

This excellent GL 4.x micro-benchmark that has been making waves recently is really interesting. Now that it's on Phoronix it's about as mainstream as it's going to get: NVIDIA Slaughters AMD Catalyst On Linux In OpenGL 4.x Micro-Benchmarks

At first glance the results sound great for NV: "The AMD Catalyst driver gets absolutely annihilated for these GL4 micro-benchmarks". But unfortunately it's bad news for everyone working in GL because it clearly demonstrates just how fractured and inconsistent the GL driver landscape actually is when the rubber hits the road.

4 comments:

  1. Gentlemen, please calm down. The first thing that would question here is: "are these tests correct?" I agree wholeheartedly with you that a this is bad news for everyone. But IMHO we should first get our base correct. I'm not an OpenGL specialist, but I've tried to find out about the apitest suite. There is really not a lot of information available about that test suite. What are the scope, aim and purpose of the tests? What is *exactly* tested? To a naive user this really looks like: "The apitest suite was developed by a Nvidia employee with the aim to destroy the competition".

    ReplyDelete
  2. The test suite is open source so anyone can dive in and check out what it's doing. Also, see the popular GDC talk "Approaching Zero Driver Overhead" here for more context: http://www.slideshare.net/CassEveritt/approaching-zero-driver-overhead

    ReplyDelete
  3. @Rich, thanks for your quick answer. Thanks for pointing to this slideshare, It nicely explains the "objective" of apitest. Maybe we should direct our flames towards phoronix, which came up with this headline: "NVIDIA Slaughters AMD Catalyst On Linux In OpenGL 4.x Micro-Benchmarks"

    ReplyDelete
  4. Phoronix is like the daily mail (or your favourite local tabloid) for Linux gaming related news with overly sensationalist headlines and otherwise shoddy content too. Most of the time it's just best to ignore it all together.

    That said, vendor B's drivers are already terrible when it comes to correctness so there wouldn't be much point in producing garbage pixels much faster. As you said in your newer post we can only hope they all one day manage to produce decent OpenGL drivers for Linux (and Windows) that work reliably.

    ReplyDelete