News ATI drivers

[Gene Heskett]

On Thursday 13 September 2007, David Bronaugh wrote:
>Gene Heskett wrote:
>> On Thursday 13 September 2007, Gabor Gombas wrote:
>>> On Thu, Sep 13, 2007 at 11:52:57AM -0400, Gene Heskett wrote:
>>>> Alan, I'm sorry to be wasting your time with this piddly-assed thread.
>>>> The point I've made, and which no one will just ignore, was that I went
>>>> out and spent something like $90 US, to buy a card that could do such as
>>>> the more recent google earth, or if I was really in a mood to waste
>>>> time, play tux racer.  That ati card, running on the radeon driver,
>>>> isn't capable of doing google earth without major breakups in the
>>>> imaging, and tux racer's response to keyboard input is so slow as to be
>>>> un-playable when the video is taking so much cpu the keyboard is ignored
>>>> for 1 to 3 seconds at a time.  I was getting <250 fps from glxgears with
>>>> it, at 95%+ cpu, now with a Jaton GForce 6200-256 and nvidia drivers I
>>>> get 1500 or so with just a few % of cpu usage.
>>>
>>> /var/log/Xorg.0.log for the machine in my office:
>>>
>>> ...
>>> (--) Chipset ATI Radeon 9200SE 5964 (AGP) found
>>> ...
>>>
>>> glxgears reports around 900 fps with ~20% CPU. I've used Google Earth a
>>> couple of times and it worked without any problems. I don't do gaming so
>>> I have no idea about tuxracer performance. The only problem is it has no
>>> DVI output, but it still produces a very nice picture on a brand-new 20"
>>> Dell flat panel.
>>>
>>> Gabor
>>
>> That's an order of magnitude better than I ever got from the card.  This
>> one was a Visiontek, with a huge ATI logo on it.
>
>Likely, direct rendering is not enabled (ie, the CPU is doing all the work).

Yes it was, and somewhere back up the log a couple of years, there is a whole 
thread related to that as you kind folks told me what to do to get it 
working.  Or at least glxinfo said it was when I'd followed all the 
instructions.  Before, glxgears ran at 55-85 frames, and the card was so slow 
I felt like I should call a surveyer & have him set some stakes.  Switching 
to another window was something I did, then went after a coffee or beer.  It 
would be switched when I got back.

>I had this stuff working 6 years ago on a Rage128, a Voodoo1, and a
>Voodoo3; later on a Radeon 7000, a Radeon 9000, a Radeon 9200, an ATI
>Mach64, an Intel i830, an S3 Savage IX, and a Via Unichrome chip (ugh).
>In each of those cases, Tuxracer and Quake3 at least would run (Google
>Earth wasn't around 6 years ago). So I think you are suffering from a
>configuration issue, not shortcomings of the driver.
>
>Perhaps you should enlist someone's help to solve the configuration
>issues, instead of accusing the devels of broken drivers.
>
>David Bronaugh

You read me wrong David.  This isn't my first dog and pony show with ATI in 
the center ring, with them taking the 5th in response to any questions.  I 
think the authors did in fact do a great job considering what they did was 
all based first on a WAG, the S optional.

I can't recall that I've ever actually seen "buggy" performance or a crash 
from the radeon driver unless you want to blame the severe pixelization and 2 
second freezes while trying to run google earth on it.  I thought then, and 
still do, that they did a great job considering the raw material they had.

The card is laying on the other table ATM, and will eventually find its way 
back into the box, and I may at some point put it in a winblows box that 
needs one.

-- 
Cheers, Gene
"There are four boxes to be used in defense of liberty:
 soap, ballot, jury, and ammo. Please use in that order."
-Ed Howdershelt (Author)
The value of a program is proportional to the weight of its output.