r300/500 rant

[bridgman]

I couldn't get the link to work... there's a site but no picture

[EmbraceUnity]

I couldn't get the link to work... there's a site but no picture

Sorry about that.

http://embraceunity.com/wp-content/uploads/2010/02/funkygears.png

It is a bit hard to see in the picture, but if you look closely there is a checkerboard effect.

[BlackStar]

GLSL 1.2 is already in place. What's missing is support for code branching, loops, and partial derivatives instructions (from the top of my head).

Yep, that's why I said GLSL 1.2 is in the works (missing dynamic flow control). Derivatives are GLSL 1.3, I think.

[pedepy]

i dont like fglrx, i just don't like that chips that are supported by catalyst also receive open source attention.

it's wasted efforts; in deciding to drop support for pre r600 chips ati more or less made the open source driver their legacy option; you don't see nvidia putting efforts in getting their legacy drivers to support brand new cards ?!

its like i said before: you either drop all catalyst efforts on linux or you stop trying to develop two drivers to certain cards wasting ressources that could be focused on bringing your legacy option up to speed with what it's intended to provide.

[bridgman]

Just to be clear, the original reasons for supporting open source driver development did not include legacy hardware support. The goals were (a) providing a great out-of-box experience for Linux users, and (b) empowering distros to provide integration and post-sale support to their customers. Not saying legacy support wasn't important, but there are a number of different ways to handle it and only some of those ways require open source drivers.

When support for the 3xx-5xx generation was dropped from fglrx we asked our developers to divert some time from the newest GPUs so that support for older GPUs could move ahead more quickly, and to fill some gaps such as power savings.

We did *not* redefine the open source graphics project to be "legacy only" and have no plans to do so in the future.

[bridgman]

One thing I don't understand -- the current 3xx-5xx drivere represent years of hard work by some very smart people, and yet you're obviously not happy with them.

If we were to stop working on new GPUs (particularly on getting initial documentation and support for new GPUs into the development community) I don't understand how there would be sufficiently high quality open source drivers by the time they were required for legacy support.

[pvtcupcakes]

i dont like fglrx, i just don't like that chips that are supported by catalyst also receive open source attention.

it's wasted efforts; in deciding to drop support for pre r600 chips ati more or less made the open source driver their legacy option; you don't see nvidia putting efforts in getting their legacy drivers to support brand new cards ?!

its like i said before: you either drop all catalyst efforts on linux or you stop trying to develop two drivers to certain cards wasting ressources that could be focused on bringing your legacy option up to speed with what it's intended to provide.

Hey man, there is no way anybody is going to pry me away from my oss radeon driver. Sure my 4850 may *work* with Catalyst, but I'm not going to touch it with a 50 foot pole.

I'd sooner go steal that Geforce 8800 that I gave to my parents.

[pingufunkybeat]

To me, fglrx is a legacy driver.

OSS drivers are the future.

[DanL]

its like i said before: you either drop all catalyst efforts on linux or you stop trying to develop two drivers to certain cards wasting ressources that could be focused on bringing your legacy option up to speed with what it's intended to provide.

I'll repeat - they're not duplicating much effort. A lot of fglrx is code from the Windows Catalyst driver and some of ATI's customers need that driver.

Reality and bottom line: Either upgrade your GPU, run a distro capable of using Catalyst 9-3, or wait patiently for the r300g driver if neither one of those options are feasible. Ranting won't help in this case. I suggest therapy or using a punching bag as more effective alternatives.

[chaos386]

One thing I don't understand -- the current 3xx-5xx drivere represent years of hard work by some very smart people, and yet you're obviously not happy with them.

If we were to stop working on new GPUs (particularly on getting initial documentation and support for new GPUs into the development community) I don't understand how there would be sufficiently high quality open source drivers by the time they were required for legacy support.

Well obviously a full-featured graphics driver for accelerated 2D, 3D, video decode, GPGPU, and power management for 3+ architectures with dozens of variants each only takes a day to code, right?