A Big Comparison Of The AMD Catalyst, Mesa & Gallium3D Drive

[agd5f]

arent they rdy for primetime like your pageflipping patches?
id love to enable pageflipping without building from git and patching the source manualls. just an xorg.conf option like color tiling

The drm pageflipping patches are scheduled to go upstream in 2.6.38, so for now you'll need to use development versions. As for the 2D tiling patches, they require 2.6.37 and probably some more stability testing before I commit them. Hopefully after I pull the ddx pflip stuff into master I'll have some time to get them integrated.

[agd5f]

With the current 1D tiling code, we only implement 1D tiling for the back buffer. Ideally we'd 2D tile all buffers and textures. r600g has preliminary support for this but it requires 2.6.37 and some ddx and r600g patches. These patches against the kms-pflip branch of the ddx:

http://people.freedesktop.org/~agd5f/all_2D_tiling.diff

and mesa:

http://people.freedesktop.org/~agd5f/2D_tiling_mesa.diff

add preliminary support.

[bridgman]

Yeah, I think we are saying the same thing. I'm going from memory so you may actually know better than me here

Either way, all of the improvements add up over time and good things are happening.

[pingufunkybeat]

I upgrade to latest git once every few days, and I seem to remember that adding the color tiling option in xorg boosted OpenArena framerate by about 60%, and other people here reported similar things.

I might have been related to a number of other optimisations which landed at the same time, I'm not sure. Bridgman should know better than I do, in any case.

I can imagine that it is more important at higher resolutions, due to the additional bandwith needed, so maybe the gain is much smaller at lower resolutions.

EDIT: you're right, I wasn't talking about averages.

[bridgman]

If you look for the largest performance gain on a single app at a single (high) resolution, I agree completely, but if we are talking about average performance gain across a range of apps (say the ones used in this article) I think the gain would be quite a bit lower.

Alex also mentioned 2D tiling, note that in some cases "1D tiling" is already supported and going from "1D tiling" to "2D tiling" (more aggressive tiling) won't necessarily bring the same gain as going from linear to 1D. Going from linear to 2D is a big deal though.

[Lynxeye]

No, I think pingufunkybeat is right.
Tiling is essential for GPU performance and should have brought a big leap in performance. 7% is way too low for propper tiling support.

[bridgman]

My recollection was that tiling on its own was good for something closer to 7%*, and there were a heap of other improvements at around the same time which collectively gave a much larger improvement. Take this with the usual disclaimer, this is just what I remember reading on the internet

* faster on some things, slower on others

[pingufunkybeat]

Well, 2D colour tiling boosted performance between 50% and 100%, which was a very significant leap.

Other optimisations might not bring such drastic improvements, but I don't expect them to be in the 1-2% range.

[agd5f]

what percentages are we talking about?
5%, 10% or more like 50%?
just rough guess.

It really depends on the workload and where the current bottlenecks are; e.g., if the app is CPU limited, hw optimizations won't make much difference. Things like tiling and hyperZ improve cache and memory bandwidth utilization, so if an app is bandwidth limited, they should help significantly.

[agd5f]

Do the docs from ATI have any information on the IL used in fglrx? You seem to get pretty good performance, so in my limited experience with compiler development, it would seem like a good idea to learn from the specialist in the domain (Thought: A generic IR with driver-specific extensions where necessary?)

AMD "IL" documentation and other shader related info is available here:

http://developer.amd.com/GPU/ATISTREAMSDK/DOCUMENTATION/Pages/default.aspx