Search:

Just a change from 5 to 4 VLIW instructions doesn't seem particularly hard to support.



They are apparently starting to work on drivers now, so an open driver could be available on launch.

Since the shader optimizer is worse than the proprietary driver and Crossfire is not supported, get the best non-59xx card that fits in your budget, case and noise/temp tolerance: ideally an HD 5870....

Thanks to Alex Deucher, Richard Li, John Bridgman, and anyone else who contributed to this effort!

Congratulations!

The work was pushed as a single commit adding about 12000 lines to the r600...

I think it should be possible to do multi-GPU rendering in a generic way (although probably not optimally).

A simple way would be to have both cards have the same data in memory.
For each...

Try to explicitly specify the hsync and vsync ranges for your monitor in xorg.conf.
The manual for the monitor may help, otherwise try random values until it works :)

Also, if the open-source...

What's the exact issue? It's unlikely that someone will be able to help if you don't specify it precisely...

If it does not compile, add "#include <linux/slab.h>" to the file that doesn't compile...

You should probably describe what the problem is exactly and report it in the fglrx bug tracker if you are sure it's a driver bug.

Also check that you are actually using the driver; if 3D is slow,...

Uh? It works fine with the Catalyst driver, just install it.

But what if your open source software is competing against proprietary software?

If you use a BSD-like license, the proprietary company can just incorporate all your code, but you can't do the...

That doesn't seem necessarily the case: r300-r700 have actively developed Gallium drivers with several community contributors (as well as the classic drivers), while the Intel Gallium drivers appear...

Isn't it enough to have the application draw to an offscreen buffer, and copy that to the compositor-accessible window with a single blit/render operation?

Assuming that the compositor syncs to...

You don't need to compile anything yourself.

Often, you can find repositories with updated components (e.g. Ubuntu PPAs) and otherwise use the packages from the development branch.

With Ubuntu,...

Sorry, I think I made a poor choice of words. I meant to say you are currently behind, not that you are not accelerating and going to fix that.

In particular, there is the basic Evergreen support,...

Good to know!

Anyway, in some sense, reverse engineering is literally asking the proprietary driver itself what you should do, and you are guaranteed to get the correct answer.

The problem, of...

I'd like to add that GPU command stream analysis is not only useful for reverse engineering, but also for tracking down open driver bugs by looking for commits changing the command stream, to compare...

By the way, note that reverse engineering is often much easier than you could think at first.

The reason is that there are tools that let you run some OpenGL program on the proprietary driver, and...

And note that the interfaces exposed by those components as a whole (the non-GPU-specific kernel ABI, OpenGL and the X11 API) are very stable and there is full commitment to keeping compatibility...

Usually you can just upgrade the kernel, libdrm, xorg and mesa to the latest stable or development versions and have the rest of the distribution work fine with no user visible changes.

Often the...

Great!

Perhaps there needs to be some equivalent of kernelnewbies.org for open GPU development: maybe Phoronix could host it and/or help with that?
Maybe it could also imitate LWN, with regular...

This is not obvious.
The problem is "just" trying to calculate faster a fixed number of iterations, and whether emulation reaches a fixed point or not seems immaterial for that.

Anyway, this is a...