Radeon 7500/8500 Support

Does anybody know about the new Radeon cards and SDL?
I am not getting the performance I expected… It may
be a problem with SDL along with XFree86 4.2.0. I can
not get any hardware support, such that I would like
to
dump frames from a V4L app onto the framebuffer… I
know direct framebuffer access under X is working
because
xawtv is making use of it. Just hoping that someone
could
shed some light on this matter. Thanks.

Shane__________________________________________________
Do You Yahoo!?
Yahoo! Greetings - Send FREE e-cards for every occasion!
http://greetings.yahoo.com

Does anybody know about the new Radeon cards and SDL?

The new Radeon cards basically don’t work very well in XFree right now.

I am not getting the performance I expected… It may
be a problem with SDL along with XFree86 4.2.0. I can
not get any hardware support, such that I would like
to
dump frames from a V4L app onto the framebuffer… I
know direct framebuffer access under X is working
because
xawtv is making use of it. Just hoping that someone
could
shed some light on this matter. Thanks.

Correct, no 3D accelleration. You can complain to XFree about this, but
they’ll ignore you. You might try Accellerated X, which does have support
for both cards. No promise it’s any good and I don’t know if they’ll let
you try it before you fork over a ridiculous amount of cash for a closed
source thing with not much more support than you get from the DRI team,
but it IS an option and they DO say they have drivers for the new Radeon
cards.

Of course, if you’re willing to deal with a closed source solution, you
probably should have bought your 3D card from a different vendor and
gotten something that actually performs in a reasonable manner. No
promises that NVIDIA is much more stable, it just happens to actually be
fast enough to use. ;)On Wed, Feb 27, 2002 at 04:31:41PM -0800, Shane Walton wrote:


Joseph Carter Don’t feed the sigs

  • james would be more impressed if netgod’s magic powers could stop the
    splits in the first place…
  • netgod notes debian developers are notoriously hard to impress

-------------- next part --------------
A non-text attachment was scrubbed…
Name: not available
Type: application/pgp-signature
Size: 273 bytes
Desc: not available
URL: http://lists.libsdl.org/pipermail/sdl-libsdl.org/attachments/20020227/db9dad80/attachment.pgp

Joseph,

Thank you for the reply. I am still somewhat in a
disagreement. If xawtv works, which uses raw XFree86
via its xvideo module, then why does SDL not just make
use of the same mechanism instead of falling back to
a system memory surface? All I want is direct 2D
performance. X 4.2.0 just doesn’t seem to be the
problem. I am going to do some digging to find out
what the differences are in the X calls. I just
didn’t
want to waste time if someone knew that SDL had a
problem with X 4.2.0. Thanks again.

Regards,

Shane M. Walton, Software Engineer
Digital System Resources, Inc.
swalton at dsrnet.com
703.234.1674

— Joseph Carter wrote:> On Wed, Feb 27, 2002 at 04:31:41PM -0800, Shane Walton wrote:

Does anybody know about the new Radeon cards and
SDL?

The new Radeon cards basically don’t work very well
in XFree right now.

I am not getting the performance I expected… It
may
be a problem with SDL along with XFree86 4.2.0. I
can
not get any hardware support, such that I would
like
to
dump frames from a V4L app onto the framebuffer…
I
know direct framebuffer access under X is working
because
xawtv is making use of it. Just hoping that
someone
could
shed some light on this matter. Thanks.

Correct, no 3D accelleration. You can complain to
XFree about this, but
they’ll ignore you. You might try Accellerated X,
which does have support
for both cards. No promise it’s any good and I
don’t know if they’ll let
you try it before you fork over a ridiculous amount
of cash for a closed
source thing with not much more support than you get
from the DRI team,
but it IS an option and they DO say they have
drivers for the new Radeon
cards.

Of course, if you’re willing to deal with a closed
source solution, you
probably should have bought your 3D card from a
different vendor and
gotten something that actually performs in a
reasonable manner. No
promises that NVIDIA is much more stable, it just
happens to actually be
fast enough to use. :wink:


Joseph Carter
Don’t feed the sigs

  • james would be more impressed if netgod’s magic
    powers could stop the
    splits in the first place…
  • netgod notes debian developers are notoriously
    hard to impress

ATTACHMENT part 2 application/pgp-signature


Do You Yahoo!?
Yahoo! Greetings - Send FREE e-cards for every occasion!
http://greetings.yahoo.com

That’s because I totally misread a few key lines of your mail. As my
NVIDIA card is actually unstable with Xvideo, I don’t use it. XawTV works
the old fashioned way here. Since you understand that XawTV uses DMA to
put the data directly into video memory, I’m unfortunately not able to be
much help otherwise.On Thu, Feb 28, 2002 at 09:30:27AM -0800, Shane Walton wrote:

Thank you for the reply. I am still somewhat in a
disagreement.


Joseph Carter You’re entitled to my opinion

<|Rain|> with sane code, maybe I could figure out the renderer :slight_smile:
rain: I’d probably be the one writing the renderer
<|Rain|> well, er, uh

-------------- next part --------------
A non-text attachment was scrubbed…
Name: not available
Type: application/pgp-signature
Size: 273 bytes
Desc: not available
URL: http://lists.libsdl.org/pipermail/sdl-libsdl.org/attachments/20020228/17d6b990/attachment.pgp

SDL doesnt care about the card you are using. It only cares about what you
are using to access the card. Fbcon, svgalib, xfree. XFree currently works
fine with any Raedon 1 board, including VE, LE, and 7500. (7500 just being
an overclocked VE). 8500 I belive 2D works fine (dont quote me on that) but
3D does not work well yet. They are working on drivers for that as we speak.On 27-Feb-2002, Shane Walton wrote:

Does anybody know about the new Radeon cards and SDL?
I am not getting the performance I expected… It may
be a problem with SDL along with XFree86 4.2.0. I can
not get any hardware support, such that I would like
to
dump frames from a V4L app onto the framebuffer… I
know direct framebuffer access under X is working
because
xawtv is making use of it. Just hoping that someone
could
shed some light on this matter. Thanks.

Shane


Do You Yahoo!?
Yahoo! Greetings - Send FREE e-cards for every occasion!
http://greetings.yahoo.com


SDL mailing list
SDL at libsdl.org
http://www.libsdl.org/mailman/listinfo/sdl


Patrick “Diablo-D3” McFarland || unknown at panax.com
"Don’t worry about how ugly you are, worry about how ugly your code is."

Er, wrong. Joe, you’re thinking 8500. 8500 doesnt work right. 7500 works fine,
its an overclocked VE. Literally. Raedon 1 chip, not the Raedon 2 chip
that are in 8500s. Also, Ill offically say, stay away from nvidia on linux.
On 27-Feb-2002, Joseph Carter wrote:

On Wed, Feb 27, 2002 at 04:31:41PM -0800, Shane Walton wrote:

Does anybody know about the new Radeon cards and SDL?

The new Radeon cards basically don’t work very well in XFree right now.

I am not getting the performance I expected… It may
be a problem with SDL along with XFree86 4.2.0. I can
not get any hardware support, such that I would like
to
dump frames from a V4L app onto the framebuffer… I
know direct framebuffer access under X is working
because
xawtv is making use of it. Just hoping that someone
could
shed some light on this matter. Thanks.

Correct, no 3D accelleration. You can complain to XFree about this, but
they’ll ignore you. You might try Accellerated X, which does have support
for both cards. No promise it’s any good and I don’t know if they’ll let
you try it before you fork over a ridiculous amount of cash for a closed
source thing with not much more support than you get from the DRI team,
but it IS an option and they DO say they have drivers for the new Radeon
cards.

Of course, if you’re willing to deal with a closed source solution, you
probably should have bought your 3D card from a different vendor and
gotten something that actually performs in a reasonable manner. No
promises that NVIDIA is much more stable, it just happens to actually be
fast enough to use. :wink:


Joseph Carter Don’t feed the sigs

  • james would be more impressed if netgod’s magic powers could stop the
    splits in the first place…
  • netgod notes debian developers are notoriously hard to impress


Patrick “Diablo-D3” McFarland || unknown at panax.com
"Don’t worry about how ugly you are, worry about how ugly your code is."

The only thing I think you might be talking about is dga, which is whacky on
quite a few cards. (Sort of like how certain cards also dont get along with
abusive directdraw applications) Xfree doesnt work on the same principals as
I think you are using. DGA isnt alot faster on high end agp cards. X (tends to)
use whatever it can to speed up display speeds. So, using a software surface
with sdl doesnt make it all that slower. At the most, you loose 5%-10% speed.On 28-Feb-2002, Shane Walton wrote:

Joseph,

Thank you for the reply. I am still somewhat in a
disagreement. If xawtv works, which uses raw XFree86
via its xvideo module, then why does SDL not just make
use of the same mechanism instead of falling back to
a system memory surface? All I want is direct 2D
performance. X 4.2.0 just doesn’t seem to be the
problem. I am going to do some digging to find out
what the differences are in the X calls. I just
didn’t
want to waste time if someone knew that SDL had a
problem with X 4.2.0. Thanks again.

Regards,

Shane M. Walton, Software Engineer
Digital System Resources, Inc.
swalton at dsrnet.com
703.234.1674

— Joseph Carter wrote:

On Wed, Feb 27, 2002 at 04:31:41PM -0800, Shane Walton wrote:

Does anybody know about the new Radeon cards and
SDL?

The new Radeon cards basically don’t work very well
in XFree right now.

I am not getting the performance I expected… It
may
be a problem with SDL along with XFree86 4.2.0. I
can
not get any hardware support, such that I would
like
to
dump frames from a V4L app onto the framebuffer…
I
know direct framebuffer access under X is working
because
xawtv is making use of it. Just hoping that
someone
could
shed some light on this matter. Thanks.

Correct, no 3D accelleration. You can complain to
XFree about this, but
they’ll ignore you. You might try Accellerated X,
which does have support
for both cards. No promise it’s any good and I
don’t know if they’ll let
you try it before you fork over a ridiculous amount
of cash for a closed
source thing with not much more support than you get
from the DRI team,
but it IS an option and they DO say they have
drivers for the new Radeon
cards.

Of course, if you’re willing to deal with a closed
source solution, you
probably should have bought your 3D card from a
different vendor and
gotten something that actually performs in a
reasonable manner. No
promises that NVIDIA is much more stable, it just
happens to actually be
fast enough to use. :wink:


Joseph Carter
Don’t feed the sigs

  • james would be more impressed if netgod’s magic
    powers could stop the
    splits in the first place…
  • netgod notes debian developers are notoriously
    hard to impress

ATTACHMENT part 2 application/pgp-signature


Do You Yahoo!?
Yahoo! Greetings - Send FREE e-cards for every occasion!
http://greetings.yahoo.com


SDL mailing list
SDL at libsdl.org
http://www.libsdl.org/mailman/listinfo/sdl


Patrick “Diablo-D3” McFarland || unknown at panax.com
"Don’t worry about how ugly you are, worry about how ugly your code is."

donning asbestos shorts

The closed source driver is Precision Insight’s fault, not NVIDIA’s. The
needed specs were offered. And the offer was rejected because the specs
NVIDIA offered wouldn’t have fit within the DRI framework. DRI assumes
that you’re always going to be writing to the bare registers. You never
do that with any of the NVIDIA cards, you use the unified architecture,
which is higher-level.

While the drivers have a number of really annoying bugs (especially Xvideo
issues), NVIDIA cannot be faulted for refusing to give up specs for their
cards. They were willing to give up exactly what they give to their own
driver programmers. That wasn’t good enough.

Clearly the DRI people have been taking lessons in negotiation from RMS.
Regardless of whether what they’re offering is good enough, if it’s not
what you want, call it unacceptable. If they cave, you win.

If they don’t, shout out to everyone that they’re an immoral corporation
who is trying to hurt free software. Most corporations cannot afford the
amount of damage to their appearance of professionalism getting into a
shouting match with someone questioning the morals of their business
practices, so they will ignore it. Of course, people automatically assume
that an unanswered accusation is probably true, so you win anyway.

I suppose there will be at least a dozen replies to this demanding me to
take back what I’ve said about the “creator” of free software, accusing me
of all sorts of heresies. All I can say is the ends do NOT justify the
means by any stretch of the imagination. But that won’t stop many of you
from flaming, so please do it off list. =pOn Thu, Feb 28, 2002 at 03:51:41PM -0500, Patrick McFarland wrote:

Er, wrong. Joe, you’re thinking 8500. 8500 doesnt work right. 7500 works fine,
its an overclocked VE. Literally. Raedon 1 chip, not the Raedon 2 chip
that are in 8500s. Also, Ill offically say, stay away from nvidia on linux.


Joseph Carter glDisable (DX8_CRAP);

“They are both businesses - if you have given them enough money, I’m
sure they’ll do whatever the hell you ask:->”
– David Welton

-------------- next part --------------
A non-text attachment was scrubbed…
Name: not available
Type: application/pgp-signature
Size: 273 bytes
Desc: not available
URL: http://lists.libsdl.org/pipermail/sdl-libsdl.org/attachments/20020228/bbbf3b34/attachment.pgp

Granted that the 7500 and 8500 are based on different
chips, I own both of these cards and they behave
exactly the same with XFree86 4.2.0 and SDL 1.2.3.
There is no hardware support. I am only getting a
software surface. Thanks.

— Joseph Carter wrote:> On Thu, Feb 28, 2002 at 03:51:41PM -0500, Patrick McFarland wrote:

Er, wrong. Joe, you’re thinking 8500. 8500 doesnt
work right. 7500 works fine,
its an overclocked VE. Literally. Raedon 1 chip,
not the Raedon 2 chip
that are in 8500s. Also, Ill offically say, stay
away from nvidia on linux.

donning asbestos shorts

The closed source driver is Precision Insight’s
fault, not NVIDIA’s. The
needed specs were offered. And the offer was
rejected because the specs
NVIDIA offered wouldn’t have fit within the DRI
framework. DRI assumes
that you’re always going to be writing to the bare
registers. You never
do that with any of the NVIDIA cards, you use the
unified architecture,
which is higher-level.

While the drivers have a number of really annoying
bugs (especially Xvideo
issues), NVIDIA cannot be faulted for refusing to
give up specs for their
cards. They were willing to give up exactly what
they give to their own
driver programmers. That wasn’t good enough.

Clearly the DRI people have been taking lessons in
negotiation from RMS.
Regardless of whether what they’re offering is good
enough, if it’s not
what you want, call it unacceptable. If they cave,
you win.

If they don’t, shout out to everyone that they’re an
immoral corporation
who is trying to hurt free software. Most
corporations cannot afford the
amount of damage to their appearance of
professionalism getting into a
shouting match with someone questioning the morals
of their business
practices, so they will ignore it. Of course,
people automatically assume
that an unanswered accusation is probably true, so
you win anyway.

I suppose there will be at least a dozen replies to
this demanding me to
take back what I’ve said about the “creator” of free
software, accusing me
of all sorts of heresies. All I can say is the ends
do NOT justify the
means by any stretch of the imagination. But that
won’t stop many of you
from flaming, so please do it off list. =p


Joseph Carter
glDisable (DX8_CRAP);

“They are both businesses - if you have given them
enough money, I’m
sure they’ll do whatever the hell you ask:->”
– David Welton

ATTACHMENT part 2 application/pgp-signature


Do You Yahoo!?
Yahoo! Greetings - Send FREE e-cards for every occasion!
http://greetings.yahoo.com

Thanks. I know X11 works fine with the 7500/8500,
that is not the issue. The issue is why does SDL not
allow direct hardware support when it is nothing more
than a layer over the OS’s graphics interface whether
it be X11, Microsoft, etc. I suppose that SDL is
broke for the time being, unless someone can explain
why I can do a memory map from a frame grabbing device
directly to the frame buffer of the Radeon 7500/8500
with X11 calls. Yet I cannot memory map directly with
SDL because it will not give me a hardware surface, it
will only give me a software surface which in turn
means I have uneeded overhead. Any explanations
welcome.

— Patrick McFarland wrote:> SDL doesnt care about the card you are using. It

only cares about what you
are using to access the card. Fbcon, svgalib, xfree.
XFree currently works
fine with any Raedon 1 board, including VE, LE, and
7500. (7500 just being
an overclocked VE). 8500 I belive 2D works fine
(dont quote me on that) but
3D does not work well yet. They are working on
drivers for that as we speak.

On 27-Feb-2002, Shane Walton wrote:

Does anybody know about the new Radeon cards and
SDL?
I am not getting the performance I expected… It
may
be a problem with SDL along with XFree86 4.2.0. I
can
not get any hardware support, such that I would
like
to
dump frames from a V4L app onto the framebuffer…
I
know direct framebuffer access under X is working
because
xawtv is making use of it. Just hoping that
someone
could
shed some light on this matter. Thanks.

Shane


Do You Yahoo!?
Yahoo! Greetings - Send FREE e-cards for every
occasion!
http://greetings.yahoo.com


SDL mailing list
SDL at libsdl.org
http://www.libsdl.org/mailman/listinfo/sdl


Patrick “Diablo-D3” McFarland || unknown at panax.com
"Don’t worry about how ugly you are, worry about how
ugly your code is."


SDL mailing list
SDL at libsdl.org
http://www.libsdl.org/mailman/listinfo/sdl


Do You Yahoo!?
Yahoo! Greetings - Send FREE e-cards for every occasion!
http://greetings.yahoo.com

Er, wrong. Joe, you’re thinking 8500. 8500 doesnt work right. 7500 works fine,
its an overclocked VE. Literally. Raedon 1 chip, not the Raedon 2 chip
that are in 8500s. Also, Ill offically say, stay away from nvidia on linux.

donning asbestos shorts

The closed source driver is Precision Insight’s fault, not NVIDIA’s. The
needed specs were offered. And the offer was rejected because the specs
NVIDIA offered wouldn’t have fit within the DRI framework. DRI assumes
that you’re always going to be writing to the bare registers. You never
do that with any of the NVIDIA cards, you use the unified architecture,
which is higher-level.

With that, I really dont care. Its nvidia’s fault for not making the drivers
open source anyhow. The Kernel crew, xfree86, and dri would gladdly ship
nvidia drivers with X and the kernel if they were open source.
e

While the drivers have a number of really annoying bugs (especially Xvideo
issues), NVIDIA cannot be faulted for refusing to give up specs for their
cards. They were willing to give up exactly what they give to their own
driver programmers. That wasn’t good enough.

Well, it costs them money. Already, the tuxbox project is dropping nvidia
hardware (was gonna use the gf2mx series, shit, but hey, its a console)
For raedon VEs or 7500s.

Clearly the DRI people have been taking lessons in negotiation from RMS.
Regardless of whether what they’re offering is good enough, if it’s not
what you want, call it unacceptable. If they cave, you win.

If they don’t, shout out to everyone that they’re an immoral corporation
who is trying to hurt free software. Most corporations cannot afford the
amount of damage to their appearance of professionalism getting into a
shouting match with someone questioning the morals of their business
practices, so they will ignore it. Of course, people automatically assume
that an unanswered accusation is probably true, so you win anyway.

I suppose there will be at least a dozen replies to this demanding me to
take back what I’ve said about the “creator” of free software, accusing me
of all sorts of heresies. All I can say is the ends do NOT justify the
means by any stretch of the imagination. But that won’t stop many of you
from flaming, so please do it off list. =p

Actually, I think rms in a flaming troll sometimes. But hey, thats him. =)
Hopefully he will gpl his dna, then we can alter it and take out the troll
dna, and then clone him using the new dna, kill -9 the original, then ./rmsOn 28-Feb-2002, Joseph Carter wrote:

On Thu, Feb 28, 2002 at 03:51:41PM -0500, Patrick McFarland wrote:


Joseph Carter glDisable (DX8_CRAP);

“They are both businesses - if you have given them enough money, I’m
sure they’ll do whatever the hell you ask:->”
– David Welton


Patrick “Diablo-D3” McFarland || unknown at panax.com
"Don’t worry about how ugly you are, worry about how ugly your code is."

The problem is, sdl always uses software stuff under X. DGA requires that the
app is either root or suid root. But with X, its pretty damn fast anyhow.On 28-Feb-2002, Shane Walton wrote:

Granted that the 7500 and 8500 are based on different
chips, I own both of these cards and they behave
exactly the same with XFree86 4.2.0 and SDL 1.2.3.
There is no hardware support. I am only getting a
software surface. Thanks.

— Joseph Carter wrote:

On Thu, Feb 28, 2002 at 03:51:41PM -0500, Patrick McFarland wrote:

Er, wrong. Joe, you’re thinking 8500. 8500 doesnt
work right. 7500 works fine,
its an overclocked VE. Literally. Raedon 1 chip,
not the Raedon 2 chip
that are in 8500s. Also, Ill offically say, stay
away from nvidia on linux.

donning asbestos shorts

The closed source driver is Precision Insight’s
fault, not NVIDIA’s. The
needed specs were offered. And the offer was
rejected because the specs
NVIDIA offered wouldn’t have fit within the DRI
framework. DRI assumes
that you’re always going to be writing to the bare
registers. You never
do that with any of the NVIDIA cards, you use the
unified architecture,
which is higher-level.

While the drivers have a number of really annoying
bugs (especially Xvideo
issues), NVIDIA cannot be faulted for refusing to
give up specs for their
cards. They were willing to give up exactly what
they give to their own
driver programmers. That wasn’t good enough.

Clearly the DRI people have been taking lessons in
negotiation from RMS.
Regardless of whether what they’re offering is good
enough, if it’s not
what you want, call it unacceptable. If they cave,
you win.

If they don’t, shout out to everyone that they’re an
immoral corporation
who is trying to hurt free software. Most
corporations cannot afford the
amount of damage to their appearance of
professionalism getting into a
shouting match with someone questioning the morals
of their business
practices, so they will ignore it. Of course,
people automatically assume
that an unanswered accusation is probably true, so
you win anyway.

I suppose there will be at least a dozen replies to
this demanding me to
take back what I’ve said about the “creator” of free
software, accusing me
of all sorts of heresies. All I can say is the ends
do NOT justify the
means by any stretch of the imagination. But that
won’t stop many of you
from flaming, so please do it off list. =p


Joseph Carter
glDisable (DX8_CRAP);

“They are both businesses - if you have given them
enough money, I’m
sure they’ll do whatever the hell you ask:->”
– David Welton

ATTACHMENT part 2 application/pgp-signature


Do You Yahoo!?
Yahoo! Greetings - Send FREE e-cards for every occasion!
http://greetings.yahoo.com


SDL mailing list
SDL at libsdl.org
http://www.libsdl.org/mailman/listinfo/sdl


Patrick “Diablo-D3” McFarland || unknown at panax.com
"Don’t worry about how ugly you are, worry about how ugly your code is."

Thanks again… My misunderstanding, I thought I had
seen direct framebuffer performace when I was using a
Matrox G450. Yes, you are right X is pretty fast when
using a software surface. I get about a 15% load on
my system when drawing 640x480 32bit @ 30Hz, which is
very impressive. The problem with using the env
SDL_VIDEODRIVER=dga etc. It consumes the whole frame
buffer so that other apps are in a sence overridden.
That is only great for full screen viewing… Oh
well, I guess raw X calls are required. Thanks again.

— Patrick McFarland wrote:> The problem is, sdl always uses software stuff under

X. DGA requires that the
app is either root or suid root. But with X, its
pretty damn fast anyhow.

On 28-Feb-2002, Shane Walton wrote:

Granted that the 7500 and 8500 are based on
different
chips, I own both of these cards and they behave
exactly the same with XFree86 4.2.0 and SDL 1.2.3.
There is no hardware support. I am only getting a
software surface. Thanks.

— Joseph Carter wrote:

On Thu, Feb 28, 2002 at 03:51:41PM -0500, Patrick McFarland wrote:

Er, wrong. Joe, you’re thinking 8500. 8500
doesnt

work right. 7500 works fine,

its an overclocked VE. Literally. Raedon 1
chip,

not the Raedon 2 chip

that are in 8500s. Also, Ill offically say,
stay

away from nvidia on linux.

<insert closed source driver and bad
performance

rant here>

donning asbestos shorts

The closed source driver is Precision Insight’s
fault, not NVIDIA’s. The
needed specs were offered. And the offer was
rejected because the specs
NVIDIA offered wouldn’t have fit within the DRI
framework. DRI assumes
that you’re always going to be writing to the
bare

registers. You never
do that with any of the NVIDIA cards, you use
the

unified architecture,
which is higher-level.

While the drivers have a number of really
annoying

bugs (especially Xvideo
issues), NVIDIA cannot be faulted for refusing
to

give up specs for their
cards. They were willing to give up exactly
what

they give to their own
driver programmers. That wasn’t good enough.

Clearly the DRI people have been taking lessons
in

negotiation from RMS.
Regardless of whether what they’re offering is
good

enough, if it’s not
what you want, call it unacceptable. If they
cave,

you win.

If they don’t, shout out to everyone that
they’re an

immoral corporation
who is trying to hurt free software. Most
corporations cannot afford the
amount of damage to their appearance of
professionalism getting into a
shouting match with someone questioning the
morals

of their business
practices, so they will ignore it. Of course,
people automatically assume
that an unanswered accusation is probably true,
so

you win anyway.

I suppose there will be at least a dozen replies
to

this demanding me to
take back what I’ve said about the “creator” of
free

software, accusing me
of all sorts of heresies. All I can say is the
ends

do NOT justify the
means by any stretch of the imagination. But
that

won’t stop many of you
from flaming, so please do it off list. =p


Joseph Carter

glDisable (DX8_CRAP);

"They are both businesses - if you have given
them

enough money, I’m
sure they’ll do whatever the hell you ask:->"
– David Welton

ATTACHMENT part 2 application/pgp-signature


Do You Yahoo!?
Yahoo! Greetings - Send FREE e-cards for every
occasion!
http://greetings.yahoo.com


SDL mailing list
SDL at libsdl.org
http://www.libsdl.org/mailman/listinfo/sdl


Patrick “Diablo-D3” McFarland || unknown at panax.com
"Don’t worry about how ugly you are, worry about how
ugly your code is."


SDL mailing list
SDL at libsdl.org
http://www.libsdl.org/mailman/listinfo/sdl


Do You Yahoo!?
Yahoo! Greetings - Send FREE e-cards for every occasion!
http://greetings.yahoo.com

I have a g400, and seeing as g400, 450, and 550 are all basically the same
chip (and same driver) No, It doesnt have direct fb stuff. And, actually,
I think dga might be broke on it too.On 28-Feb-2002, Shane Walton wrote:

Thanks again… My misunderstanding, I thought I had
seen direct framebuffer performace when I was using a
Matrox G450. Yes, you are right X is pretty fast when
using a software surface. I get about a 15% load on
my system when drawing 640x480 32bit @ 30Hz, which is
very impressive. The problem with using the env
SDL_VIDEODRIVER=dga etc. It consumes the whole frame
buffer so that other apps are in a sence overridden.
That is only great for full screen viewing… Oh
well, I guess raw X calls are required. Thanks again.

— Patrick McFarland wrote:

The problem is, sdl always uses software stuff under
X. DGA requires that the
app is either root or suid root. But with X, its
pretty damn fast anyhow.

On 28-Feb-2002, Shane Walton wrote:

Granted that the 7500 and 8500 are based on
different
chips, I own both of these cards and they behave
exactly the same with XFree86 4.2.0 and SDL 1.2.3.
There is no hardware support. I am only getting a
software surface. Thanks.

— Joseph Carter wrote:

On Thu, Feb 28, 2002 at 03:51:41PM -0500, Patrick McFarland wrote:

Er, wrong. Joe, you’re thinking 8500. 8500
doesnt

work right. 7500 works fine,

its an overclocked VE. Literally. Raedon 1
chip,

not the Raedon 2 chip

that are in 8500s. Also, Ill offically say,
stay

away from nvidia on linux.

<insert closed source driver and bad
performance

rant here>

donning asbestos shorts

The closed source driver is Precision Insight’s
fault, not NVIDIA’s. The
needed specs were offered. And the offer was
rejected because the specs
NVIDIA offered wouldn’t have fit within the DRI
framework. DRI assumes
that you’re always going to be writing to the
bare

registers. You never
do that with any of the NVIDIA cards, you use
the

unified architecture,
which is higher-level.

While the drivers have a number of really
annoying

bugs (especially Xvideo
issues), NVIDIA cannot be faulted for refusing
to

give up specs for their
cards. They were willing to give up exactly
what

they give to their own
driver programmers. That wasn’t good enough.

Clearly the DRI people have been taking lessons
in

negotiation from RMS.
Regardless of whether what they’re offering is
good

enough, if it’s not
what you want, call it unacceptable. If they
cave,

you win.

If they don’t, shout out to everyone that
they’re an

immoral corporation
who is trying to hurt free software. Most
corporations cannot afford the
amount of damage to their appearance of
professionalism getting into a
shouting match with someone questioning the
morals

of their business
practices, so they will ignore it. Of course,
people automatically assume
that an unanswered accusation is probably true,
so

you win anyway.

I suppose there will be at least a dozen replies
to

this demanding me to
take back what I’ve said about the “creator” of
free

software, accusing me
of all sorts of heresies. All I can say is the
ends

do NOT justify the
means by any stretch of the imagination. But
that

won’t stop many of you
from flaming, so please do it off list. =p


Joseph Carter

glDisable (DX8_CRAP);

"They are both businesses - if you have given
them

enough money, I’m
sure they’ll do whatever the hell you ask:->"
– David Welton

ATTACHMENT part 2 application/pgp-signature


Do You Yahoo!?
Yahoo! Greetings - Send FREE e-cards for every
occasion!
http://greetings.yahoo.com


SDL mailing list
SDL at libsdl.org
http://www.libsdl.org/mailman/listinfo/sdl


Patrick “Diablo-D3” McFarland || unknown at panax.com
"Don’t worry about how ugly you are, worry about how
ugly your code is."


SDL mailing list
SDL at libsdl.org
http://www.libsdl.org/mailman/listinfo/sdl


Do You Yahoo!?
Yahoo! Greetings - Send FREE e-cards for every occasion!
http://greetings.yahoo.com


SDL mailing list
SDL at libsdl.org
http://www.libsdl.org/mailman/listinfo/sdl


Patrick “Diablo-D3” McFarland || unknown at panax.com
"Don’t worry about how ugly you are, worry about how ugly your code is."

Er, wrong. Joe, you’re thinking 8500. 8500 doesnt work right. 7500 works fine,
its an overclocked VE.

Actually not... The VE is basically a radeon '1' (rv100/rg6),

without TCL, 2 CRTC’s, and only a single pipeline.

The 7500 is a shrink-die radeon '1' with faster clocks and 2

CRTC’s.

The 8500 is based on the rv200 chip.  2D-wise, very similiar to

the rv100. 3D-wise, just a tad different ;-)On Thu, 28 Feb 2002, Patrick McFarland wrote:

Date: Thu, 28 Feb 2002 15:51:41 -0500
From: Patrick McFarland
Reply-To: sdl at libsdl.org
To: sdl at libsdl.org
Subject: Re: [SDL] Radeon 7500/8500 Support

Literally. Raedon 1 chip, not the Raedon 2 chip
that are in 8500s. Also, Ill offically say, stay away from nvidia on linux.

On 27-Feb-2002, Joseph Carter wrote:

On Wed, Feb 27, 2002 at 04:31:41PM -0800, Shane Walton wrote:

Does anybody know about the new Radeon cards and SDL?

The new Radeon cards basically don’t work very well in XFree right now.

I am not getting the performance I expected… It may
be a problem with SDL along with XFree86 4.2.0. I can
not get any hardware support, such that I would like
to
dump frames from a V4L app onto the framebuffer… I
know direct framebuffer access under X is working
because
xawtv is making use of it. Just hoping that someone
could
shed some light on this matter. Thanks.

Correct, no 3D accelleration. You can complain to XFree about this, but
they’ll ignore you. You might try Accellerated X, which does have support
for both cards. No promise it’s any good and I don’t know if they’ll let
you try it before you fork over a ridiculous amount of cash for a closed
source thing with not much more support than you get from the DRI team,
but it IS an option and they DO say they have drivers for the new Radeon
cards.

Of course, if you’re willing to deal with a closed source solution, you
probably should have bought your 3D card from a different vendor and
gotten something that actually performs in a reasonable manner. No
promises that NVIDIA is much more stable, it just happens to actually be
fast enough to use. :wink:


Joseph Carter Don’t feed the sigs

  • james would be more impressed if netgod’s magic powers could stop the
    splits in the first place…
  • netgod notes debian developers are notoriously hard to impress


Patrick “Diablo-D3” McFarland || unknown at panax.com
"Don’t worry about how ugly you are, worry about how ugly your code is."


SDL mailing list
SDL at libsdl.org
http://www.libsdl.org/mailman/listinfo/sdl


Jon Trulson mailto:@Jon_Trulson
ID: 1A9A2B09, FP: C23F328A721264E7 B6188192EC733962
PGP keys at http://radscan.com/~jon/PGPKeys.txt
#include <std/disclaimer.h>
Bad Color Temperature, Too much Peach.

Er, wrong. Joe, you’re thinking 8500. 8500 doesnt work right. 7500 works fine,
its an overclocked VE.

Actually not… The VE is basically a radeon ‘1’ (rv100/rg6),
without TCL, 2 CRTC’s, and only a single pipeline.

The 7500 is a shrink-die radeon ‘1’ with faster clocks and 2
CRTC’s.

I was misinformed over this point then. Performance wise, how more efficient
is the 7500 over the VE?On 28-Feb-2002, Jon Trulson wrote:

On Thu, 28 Feb 2002, Patrick McFarland wrote:

Date: Thu, 28 Feb 2002 15:51:41 -0500
From: Patrick McFarland
Reply-To: sdl at libsdl.org
To: sdl at libsdl.org
Subject: Re: [SDL] Radeon 7500/8500 Support

The 8500 is based on the rv200 chip. 2D-wise, very similiar to
the rv100. 3D-wise, just a tad different :wink:

Literally. Raedon 1 chip, not the Raedon 2 chip
that are in 8500s. Also, Ill offically say, stay away from nvidia on linux.

On 27-Feb-2002, Joseph Carter wrote:

On Wed, Feb 27, 2002 at 04:31:41PM -0800, Shane Walton wrote:

Does anybody know about the new Radeon cards and SDL?

The new Radeon cards basically don’t work very well in XFree right now.

I am not getting the performance I expected… It may
be a problem with SDL along with XFree86 4.2.0. I can
not get any hardware support, such that I would like
to
dump frames from a V4L app onto the framebuffer… I
know direct framebuffer access under X is working
because
xawtv is making use of it. Just hoping that someone
could
shed some light on this matter. Thanks.

Correct, no 3D accelleration. You can complain to XFree about this, but
they’ll ignore you. You might try Accellerated X, which does have support
for both cards. No promise it’s any good and I don’t know if they’ll let
you try it before you fork over a ridiculous amount of cash for a closed
source thing with not much more support than you get from the DRI team,
but it IS an option and they DO say they have drivers for the new Radeon
cards.

Of course, if you’re willing to deal with a closed source solution, you
probably should have bought your 3D card from a different vendor and
gotten something that actually performs in a reasonable manner. No
promises that NVIDIA is much more stable, it just happens to actually be
fast enough to use. :wink:


Joseph Carter Don’t feed the sigs

  • james would be more impressed if netgod’s magic powers could stop the
    splits in the first place…
  • netgod notes debian developers are notoriously hard to impress


Patrick “Diablo-D3” McFarland || unknown at panax.com
"Don’t worry about how ugly you are, worry about how ugly your code is."


SDL mailing list
SDL at libsdl.org
http://www.libsdl.org/mailman/listinfo/sdl


Jon Trulson mailto:jon at radscan.com
ID: 1A9A2B09, FP: C23F328A721264E7 B6188192EC733962
PGP keys at http://radscan.com/~jon/PGPKeys.txt
#include <std/disclaimer.h>
Bad Color Temperature, Too much Peach.


SDL mailing list
SDL at libsdl.org
http://www.libsdl.org/mailman/listinfo/sdl


Patrick “Diablo-D3” McFarland || unknown at panax.com
"Don’t worry about how ugly you are, worry about how ugly your code is."

Er, wrong. Joe, you’re thinking 8500. 8500 doesnt work right. 7500 works fine,
its an overclocked VE.

Actually not... The VE is basically a radeon '1' (rv100/rg6),

without TCL, 2 CRTC’s, and only a single pipeline.

The 7500 is a shrink-die radeon '1' with faster clocks and 2

CRTC’s.

I was misinformed over this point then. Performance wise, how more efficient
is the 7500 over the VE?

Well, that kind of depends on what benchmark(s) you use ;-)... on

viewperf benchmarks, the 7500 is anywhere from 3-6 times faster than the
VE… With q3 it’s about 2-3 times faster. x11perf (2D), about 2x.

The VE is is about on par with the G450...On Fri, 1 Mar 2002, Patrick McFarland wrote:

Date: Fri, 1 Mar 2002 00:52:48 -0500
From: Patrick McFarland
Reply-To: sdl at libsdl.org
To: sdl at libsdl.org
Subject: Re: [SDL] Radeon 7500/8500 Support
On 28-Feb-2002, Jon Trulson wrote:

On Thu, 28 Feb 2002, Patrick McFarland wrote:

Date: Thu, 28 Feb 2002 15:51:41 -0500
From: Patrick McFarland
Reply-To: sdl at libsdl.org
To: sdl at libsdl.org
Subject: Re: [SDL] Radeon 7500/8500 Support

The 8500 is based on the rv200 chip.  2D-wise, very similiar to

the rv100. 3D-wise, just a tad different :wink:

Literally. Raedon 1 chip, not the Raedon 2 chip
that are in 8500s. Also, Ill offically say, stay away from nvidia on linux.

On 27-Feb-2002, Joseph Carter wrote:

On Wed, Feb 27, 2002 at 04:31:41PM -0800, Shane Walton wrote:

Does anybody know about the new Radeon cards and SDL?

The new Radeon cards basically don’t work very well in XFree right now.

I am not getting the performance I expected… It may
be a problem with SDL along with XFree86 4.2.0. I can
not get any hardware support, such that I would like
to
dump frames from a V4L app onto the framebuffer… I
know direct framebuffer access under X is working
because
xawtv is making use of it. Just hoping that someone
could
shed some light on this matter. Thanks.

Correct, no 3D accelleration. You can complain to XFree about this, but
they’ll ignore you. You might try Accellerated X, which does have support
for both cards. No promise it’s any good and I don’t know if they’ll let
you try it before you fork over a ridiculous amount of cash for a closed
source thing with not much more support than you get from the DRI team,
but it IS an option and they DO say they have drivers for the new Radeon
cards.

Of course, if you’re willing to deal with a closed source solution, you
probably should have bought your 3D card from a different vendor and
gotten something that actually performs in a reasonable manner. No
promises that NVIDIA is much more stable, it just happens to actually be
fast enough to use. :wink:


Joseph Carter Don’t feed the sigs

  • james would be more impressed if netgod’s magic powers could stop the
    splits in the first place…
  • netgod notes debian developers are notoriously hard to impress


Patrick “Diablo-D3” McFarland || unknown at panax.com
"Don’t worry about how ugly you are, worry about how ugly your code is."


SDL mailing list
SDL at libsdl.org
http://www.libsdl.org/mailman/listinfo/sdl


Jon Trulson mailto:@Jon_Trulson
ID: 1A9A2B09, FP: C23F328A721264E7 B6188192EC733962
PGP keys at http://radscan.com/~jon/PGPKeys.txt
#include <std/disclaimer.h>
Bad Color Temperature, Too much Peach.


SDL mailing list
SDL at libsdl.org
http://www.libsdl.org/mailman/listinfo/sdl


Patrick “Diablo-D3” McFarland || unknown at panax.com
"Don’t worry about how ugly you are, worry about how ugly your code is."


SDL mailing list
SDL at libsdl.org
http://www.libsdl.org/mailman/listinfo/sdl


Jon Trulson mailto:@Jon_Trulson
ID: 1A9A2B09, FP: C23F328A721264E7 B6188192EC733962
PGP keys at http://radscan.com/~jon/PGPKeys.txt
#include <std/disclaimer.h>
Bad Color Temperature, Too much Peach.

Does anybody know about the new Radeon cards and SDL?
I am not getting the performance I expected… It may
be a problem with SDL along with XFree86 4.2.0. I can
not get any hardware support, such that I would like
to
dump frames from a V4L app onto the framebuffer… I
know direct framebuffer access under X is working
because
xawtv is making use of it. Just hoping that someone
could
shed some light on this matter. Thanks.

Let me clear this up.

xawtv is using the XVideo extension, which allows you to display YUV
image data (MPEG-video format) using hardware acceleration. However,
this feature does not allow you to do normal 2D RGB acceleration.

SDL does take advantage of the XVideo extension if you use SDL’s
YUV overlay API. But again, this isn’t useful for normal RGB display.

And finally… everybody please stop ranting about close/open source drivers on Linux! Good grief. :slight_smile:

-Sam Lantinga, Software Engineer, Blizzard Entertainment

With that, I really dont care. Its nvidia’s fault for not making the drivers
open source anyhow. The Kernel crew, xfree86, and dri would gladdly ship
nvidia drivers with X and the kernel if they were open source.

They have license agreements with SGI over parts of their drivers which do
not allow that. If you were aware of the 3dfx situation, they were unable
to release one iota of specs or code for the Voodoo Rush card (2+3D Voodoo
Graphics) because they didn’t own the IP for the 2D part. Nobody bothered
to try and figure out how it worked and modify the Voodoo Graphics drivers
to work with it. No surprise there, who still has either card? Voodoo2
cards are practically free nowadays and have been since the time of the
spec release.

While the drivers have a number of really annoying bugs (especially Xvideo
issues), NVIDIA cannot be faulted for refusing to give up specs for their
cards. They were willing to give up exactly what they give to their own
driver programmers. That wasn’t good enough.

Well, it costs them money. Already, the tuxbox project is dropping nvidia
hardware (was gonna use the gf2mx series, shit, but hey, its a console)
For raedon VEs or 7500s.

That’s too bad. The Radeon really has crappy feature support in Linux
right now. To this day, there is still no support for T&L. While none of
us care much for the &L part (OpenGL light sucks, no matter how fast you
make it!), the lack of hardware transforms will really hurt a lot of games
such as Tribes 2 and I think also RtCW. Granted those aren’t likely to be
huge games on the Tuxbox since you don’t normally think about keyboards
and mice at the TeeVee, but if the intent is to put out something that can
have quality games like those found on the Playstation 2 or Xbox, the 7500
just isn’t going to cut it by a long shot.

Actually, I think rms in a flaming troll sometimes. But hey, thats him. =)
Hopefully he will gpl his dna, then we can alter it and take out the troll
dna, and then clone him using the new dna, kill -9 the original, then ./rms

Some have argued the same about me. :wink: I certainly have no problems
saying something I expect will not make me very popular, especially if I’m
aware that others agree, but don’t want to say anything…

The GPL is basically a good license. Richard’s goals are in line with
mine until they clash with my sense of a need to choose free software over
the alternatives rather than have that choice made for me. And, as I’ve
already indicated, I don’t like his negotiating tactics. But that doesn’t
detract from the overall idea that free software and open standards are
good things. That’s why many of us are here, isn’t it? SDL is quickly
becoming one of those open standards. (Whew! Nice topic save!)On Thu, Feb 28, 2002 at 06:06:59PM -0500, Patrick McFarland wrote:


Joseph Carter Sooner or later, BOOM!

That reminds me, we’ll need to buy a chainsaw for the office. “In
case of emergency, break glass”

-------------- next part --------------
A non-text attachment was scrubbed…
Name: not available
Type: application/pgp-signature
Size: 273 bytes
Desc: not available
URL: http://lists.libsdl.org/pipermail/sdl-libsdl.org/attachments/20020301/533d85de/attachment.pgp

Hmm, now, just to doubly make sure, the 7500 does infact work with current
"raedon 1" drivers, right?–
Patrick “Diablo-D3” McFarland || unknown at panax.com
"Don’t worry about how ugly you are, worry about how ugly your code is."