Xonotic Forums

I've been doing a bit of work on the-big-benchmark.sh and attach a patch of a few changes for conveniece, none of them make a difference to the results. I'm sure these can be ported to the .bat file too. Apologies for naming the patch as a .txt, no I didn't write it in Notepad, the forum just won't let you upload a .patch file.

Changes:
- Print 'Benchmarking on $e' to the terminal to give a bit more feedback on progress
- Print out benchmark MED results to the terminal after each quality run to give a bit more feedback and give something very easy to copy to the forum
- Split the quality settings up and made it so that before Normal and Ultra it scans through the engine logs for OpenGL 2.0 not beeing found or vid_soft being enabled respectively. It then exits if these tests aren't supported so should make process simpler and much quicker for non-experts to gain more reliable results

Please use as you see fit, even if it doesn't get in the codebase it still may help a few people out when benchmarking as you can apply it to your own install. Be aware though for git or rsync users that updating will overwrite this patch!

Enjoy.

+cl_playerdetailreduction 0 really MUST stay, or the results get incompatible.

Oops, missed something when making the patch, that wasn't supposed to get there! Fixed the patch now so that it's back to how it was supposed to be.

I applied the other changes in git, and also ported them to the .bat. Not tested yet though.

Thanks div. Do you think there would be any (borderline trivial) difference if the FPS counter was enabled? I had wondered about this before and I know any difference would be difficult to find but having the FPS counter on by default in a benchmark seems logical to me. Gives you something to look at. ;-)

Yes, this also changes results. However, it would be a good idea to turn that on in the next version of the benchmark.

Such a version maybe would be a good idea soon anyway, as we then enable LOD too.