Working smarter not harder

Day 1 - Mesa development ARB_arrays_of_arrays

Day 1 is over and I've made some good progress on the piglit tests. I have a large number of compile tests done 77 to be exact although a number of them are just slight variations on each other so don't expect me to me to continue at this rate. To give a quick explanation of what I mean by compile tests basically we just test that a shader passes the compilation process at this stage there is no testing done to check if the shader was actually compiled correctly we just want to make sure that correct shader code passes and invalid code is rejected.

 

You can take a look at the tests here: https://github.com/tarceri/Mesa_arrays_of_arrays_piglit/commits/master

 

Also as I mentioned in my IndieGoGo update I'm looking for someone with a Nvidia card to help test my tests against the binary drivers. As I've mentioned before I only have AMD and Intel hardware and the AMD binary drivers don't seem to follow the spec very well. My tests fail a few times on the AMD Catalyst drivers and in one instance even cause a crash. I'm pretty sure I have everything right and these are issues with the AMD drivers but I would just like for someone to run them on the Nvidia drivers before I submit them to the piglit mailing list.

If you are interested in helping you can get my tests by cloning my repo on github: https://github.com/tarceri/Mesa_arrays_of_arrays_piglit

 

Building piglit with the instructions in the readme file, and running the following command from the piglit directory:

./piglit-run.py -t ARB_arrays_of_arrays/compiler* tests/all.tests results

 

Then if you could copy the contents of the text file called 'main' in the 'results' directory and create a new issue on github: https://github.com/tarceri/Mesa_arrays_of_arrays_piglit/issues and paste the contents there that would be awesome.

 

Plans for tomorrow:

 

At the moment my plans for tomorrow are to double check I haven't missed any major compile tests, fix any issues found if someone was able to help test, and then submit my compile tests to the piglit mailing list. Then I'm going to have a bit of fun working on the actual ARB_arrays_of_arrays extension. I've been working hard in my spare time over the past few weeks and currently have a much cleaned up work in progress branch here: https://github.com/tarceri/Mesa_arrays_of_arrays/tree/wip5

This branch now compiles with no regressions and adds some infrastructure for supporting arrays of arrays my next aim is write a very basic execution test and to get this compiling and running correctly.

No Comments Yet

Leave a Reply

Your email address will not be published. Required fields are marked *