On Fri, 2004-01-09 at 14:40, Alex Deucher wrote:
> er... well, the bitmap descriptor speed up in my last email was a
> fluke... that code I changed never actually gets run, so the speed up
> was just a random fluke I guess.
>
> What is SavageInitBuffers() in savage_dri.c supposed to do? all of it
Well I'll be dammned!!!
It was the sound! I stared up quake3 without audio, and everything ran fine.
Sorry for assuming it was a graphics problem.
Keep up the good work guys!
On Friday 09 January 2004 09:05, Vladimir Dergachev wrote:
> Have you checked whether there is some sort of sound daemon
Alan, perhaps you can structure it similarly to the XvMC support in the
savage driver. it uses the DRM similarly to the 3D driver.
Alex
--- Alan Cox <[EMAIL PROTECTED]> wrote:
> Ok I am now happy with the testing on the VIA CLE266 driver. Would
> someone care to pick up the dri module for it fro
On Fri, 9 Jan 2004 16:00:46 -0800 (PST)
Alex Deucher <[EMAIL PROTECTED]> wrote:
> whoops. I'll have to check that out. Can you point me to the code?
> I'm assuming savage_span in the DRI somewhere? Also, only 16 bpp works
The span stuff itself is in savagespan.c in the Mesa driver. However, i
whoops. I'll have to check that out. Can you point me to the code?
I'm assuming savage_span in the DRI somewhere? Also, only 16 bpp works
at the moment on savage4. I've only been able to run one 3d app at 32
bpp, glxheads from the mesa demos. all the rest spit out this error
message:
savagedm
I just noticed that your changes broke the line pitch in the span
functions (again). You see the effect when you run xscreensaver hacks with
-fps. I'll see if I can fix it once more. ;-)
Felix
On Fri, 9 Jan 2004 12:45:10 -0800 (PST)
Alex Deucher <[EMAIL PROTECTED]> wrote:
> The attached patch is
Ok I am now happy with the testing on the VIA CLE266 driver. Would
someone care to pick up the dri module for it from
http://people.redhat.com/~alan/
The 2D support is already in XFree 4.3.99 CVS. The 3D code is a port of
the VIA provided 3D code with their extensions to core mesa removed
(they
er... well, the bitmap descriptor speed up in my last email was a
fluke... that code I changed never actually gets run, so the speed up
was just a random fluke I guess.
What is SavageInitBuffers() in savage_dri.c supposed to do? all of its
content is commented out. it seems to contain the actual
I tested your patch on my ProSavage. It works and it speeds up glxgears
by over 50fps. I was surprised to read numbers over 500fps on Savage4 as
that's about what I'm getting with ProSavage too. It used to be about
490, with your patch it's about 550 now.
BTW, I just saw your other mail about the
just an FYI, I switched the code in savage_dri.c to use the global
bitmap descripter for the front buffer, the primary BD for the back
buffer and the secondary BD for the depth buffer. Previously they were
all sharing the Primary BD. that provided another 10-12 fps in
glxgears. unfortunately, it
Andi Kleen <[EMAIL PROTECTED]> wrote:
>
> > I know what happened. 2.6.1-rc2-mm1 added the `-msoft-float' compiler
> > option. So what used to be inline 387 instructions became fp library
> > calls.
> >
> > So it's not an _urgent_ problem, and I can probably just drop the
> > -msoft-float patch fr
Am Freitag, 09. Januar 2004 19:51 schrieb Brian Paul:
> Olivier Chapuis wrote:
> > Hello,
> >
> > I use "current cvs" from 2003-12-17 (dri - Mesa-newtree) on an Ati
> > Mobility M7 LW. When dri is enabled (OpenGL renderer string: Mesa DRI
> > Radeon 20030328 AGP 4xTCL) I get a memory leak when I lo
The attached patch is a cleaned up version of the one I made this
morning. It enables DRI support for savage4 and shouldn't break
prosavage or twister, however, I'd like someone with prosvage or
twister to check it out before I commit it.
Both tiled and linear framebuffer mode now work with 3D.
Olivier Chapuis wrote:
Hello,
I use "current cvs" from 2003-12-17 (dri - Mesa-newtree) on an Ati
Mobility M7 LW. When dri is enabled (OpenGL renderer string: Mesa DRI
Radeon 20030328 AGP 4xTCL) I get a memory leak when I load and unload
a texture.
Attached an example program. Put it in Mesa-newtre
The attached patch is a pretty raw. it will break prosavage and it
needs a major clean up, but for those looking to run 3D on savage4
cards, have at it. glxgears works flawlessly (~570 fps). tuxracer and
chromium BSU look a bit like someone ran an impressionist gimp filter
on them and they also l
FYI, I just got 3D and 2D working on savage4. the PBD setup was wrong
in savage_dri.c. I'll be posting a patch soon. Also, what bpp should
depth buffer be? the same as the framebuffer?
Alex
--- Felix Kühling <[EMAIL PROTECTED]> wrote:
> On Thu, 8 Jan 2004 15:40:55 -0800 (PST)
> Alex Deucher <
Am Donnerstag, 08. Januar 2004 00:41 schrieb Felix Kühling:
> Pavel,
>
> after reading your patch more carfully I decided to write the number
> parsing functions myself. No offense, but your version does not parse
> numbers in scientific notation like 1.5e-8, it is not good enough at
> detecting er
Michel DÃnzer wrote:
I've no idea if the order is really important or if this just hides
another problem, maybe someone with databooks could answer that.
I don't remember seeing any clear information about this in the docs I
have, but your patch is proof enough for me that it does matter. :)
Y
savage4 uses 8 for bci enable while it's 0 on prosavage. that's why
you have to define SAVAGE4 to make it work. the enablemode_m7()
function uses 8 arlready. the BD's and BCI seems to work differently
on savage4 vs. prosavage.
Also FWIW, I tried both with and without the cob enabled and it does
--- Felix Kühling <[EMAIL PROTECTED]> wrote:
> On Thu, 8 Jan 2004 15:40:55 -0800 (PST)
> Alex Deucher <[EMAIL PROTECTED]> wrote:
>
> > well, I managed to fix the 2D corruption in tile mode on savage4.
> the
> > problem is that the bits of the GBD (MM816C) have different meaning
> on
> > savage4
Andrew Morton wrote:
Linus Torvalds <[EMAIL PROTECTED]> wrote:
On Thu, 8 Jan 2004, Andrew Morton wrote:
Well sisfb_do_set_var() is using floating point. I guess it has been
changed in the recent update so the compiler now has to actually emit FP
library function
On Fri, 2004-01-09 at 01:47, Roland Scheidegger wrote:
> Based on the idea of Felix KÃhling that the order in which state atoms
> are emitted might be important on the radeon, I thought maybe it might
> be important on the r200 as well, even though there are no lockups.
> With the attached patch
At 10:46 AM 9/1/2004, Felix Kühling wrote:
Are you sure that tiled mode is enabled correctly? I think Rafael
reported earlier that 3D (glxgears) worked ok for him with tiled mode
enabled, even though 2D was garbled.
Yes, that's correct but after editing savage_bci.h and defining "SAVAGE4 1"
bye.
Yes you're right, I just used fixed numbers matching those in
r200_hw_state in r200_context.h. (Note this will also try to emit mtl[1]
and pix[2-5] which is unncessary, as they are never ALLOC_STATE'd nor
used, but I guess it shouldn't hurt as they'll never have their dirty
flag set.)
Attached
On Thu, 8 Jan 2004 15:40:55 -0800 (PST)
Alex Deucher <[EMAIL PROTECTED]> wrote:
> well, I managed to fix the 2D corruption in tile mode on savage4. the
> problem is that the bits of the GBD (MM816C) have different meaning on
> savage4 vs. twister/prosavage. The fix is to use the
> SavageEnableMo
I think Roland's patch can't cope with more than two texture units. As I
experimented with this while I had Andreas Stenglein's 3TMU patch
applied some of the time I decided to make something that can handle a
variable number of texture units. Since the number of texture units is
going to be config
Roland Scheidegger wrote:
Based on the idea of Felix Kühling that the order in which state atoms
are emitted might be important on the radeon, I thought maybe it might
be important on the r200 as well, even though there are no lockups.
With the attached patch indeed the random color flashings see
Have you checked whether there is some sort of sound daemon running that
prevents access to your sound card ?
best
Vladimir Dergachev
On Thu, 1 Jan 2004, MichaelM wrote:
> I tried installing the new ati binary drivers, and every game, except
28 matches
Mail list logo
