Amiga.org

Amiga computer related discussion => Amiga community support ideas => Topic started by: JohnFante on September 11, 2013, 02:54:34 PM

Title: Bounty for better Radeon Mediator drivers
Post by: JohnFante on September 11, 2013, 02:54:34 PM
Compared to the Voodoo drivers the Radeon drivers for OS.39 suck - to put it bluntly. They lack speed and there is no Warp3D support.

On OS4.1 Classic it is the other way around. The Voodoo drivers are slow and Warp3D support is not good (I can for instance not run Quake2 on my Voodoo 5 in OS4.1 Classic).

I am considering setting up a bounty to collect enough money to get Elbox to develop and release better Radeon drivers with Warp3D support. And to get the drivers open sourced :-)

I have not contacted Elbox yet to see if this is a possibility.

Before I do I just wanted to test if anybody here would be interesting in supporting such a bounty. If so it would be great. If not - no problem :-).

I know Elbox should be doing this themselves but sometimes things needs a little lubrication .... ;-)
Title: Re: Bounty for better Radeon Mediator drivers
Post by: nicholas on September 11, 2013, 03:12:57 PM
You might want to get in touch with Ratte as he was working on a new 68k Radeon driver in the past.
https://www.youtube.com/watch?v=HWax2_r0cDg
Title: Re: Bounty for better Radeon Mediator drivers
Post by: Cosmos on September 11, 2013, 03:45:58 PM
Good idea ! And think about asking a pci.library 68k romable...
Title: Re: Bounty for better Radeon Mediator drivers
Post by: JohnFante on September 11, 2013, 04:26:42 PM
Quote
You might want to get in touch with Ratte as he was working on a new 68k Radeon driver in the past.
https://www.youtube.com/watch?v=HWax2_r0cDg


As far as I know Ratte was stopped due to some copyright discussion with Elbox. :-(

Quote
Good idea ! And think about asking a pci.library 68k romable...


Great idea. I will put that on the todo-ist!
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 11, 2013, 04:33:23 PM
elbox has not developed warp3d drivers neither for radeon nor for voodoo, so they cannot do anything about it, no matter bounty or not. the w3d drivers supplied for mediator  were developed by hyperion, paid and comissioned by elbox as far as ive been told by them. as it has been stated, hyperion has no interest developing anything for amiga (68k) anymore, so i guess the case is lost.

what i imagine could be done about it (given there is someone with enough knowledge and dedication) is:
1. to develop hardware accelerated backend for wazp3d.
2. or to support aros(68k), introduce pci bridges support to it, make gallium work on it, and introduce wazp3d to gallium wrapper. sounds complex and im not sure it would actually work, but thats what i would try. mind you i have pursued this subject few years ago and turned to aros as result anyway.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: nicholas on September 11, 2013, 05:56:36 PM
Quote from: JohnFante;747605
As far as I know Ratte was stopped due to some copyright discussion with Elbox. :-(


True but there is nothing legally stopping him for providing a binary diff between his driver and Elbox' and then end users patching their own version with his diff. ;)
Title: Re: Bounty for better Radeon Mediator drivers
Post by: Oldsmobile_Mike on September 11, 2013, 06:25:49 PM
Quote from: JohnFante;747605
As far as I know Ratte was stopped due to some copyright discussion with Elbox. :-(


Lovely!  Let's see, what would benefit Amiga users and the community more?  A new product, or more backbiting, lawsuits, and in-fighting?  Oh, wait.  :(  :(  :(  :(  /rant
Title: Re: Bounty for better Radeon Mediator drivers
Post by: JohnFante on September 11, 2013, 07:02:26 PM
Quote from: wawrzon;747606
elbox has not developed warp3d drivers neither for radeon nor for voodoo, so they cannot do anything about it, no matter bounty or not. the w3d drivers supplied for mediator  were developed by hyperion, paid and comissioned by elbox as far as ive been told by them. as it has been stated, hyperion has no interest developing anything for amiga (68k) anymore, so i guess the case is lost.


Maybe they could be convinced to opensource Warp3D then. That is also the only solution that has any kind of an future.

Just trying to sense the mood - and find a way :-)
Title: Re: Bounty for better Radeon Mediator drivers
Post by: nicholas on September 11, 2013, 07:04:21 PM
Quote from: JohnFante;747625
Maybe they could be convinced to opensource Warp3D then. That is also the only solution that has any kind of an future.

Just trying to sense the mood - and find a way :-)

That will never happen. Hyperion (and some of their contractors) has proved by its past actions to be one of the most anti-opensource companies I've ever seen.

HW accelerated back end to Wazp3D is the best bet.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 11, 2013, 07:50:23 PM
Quote from: JohnFante;747625
Maybe they could be convinced to opensource Warp3D then. That is also the only solution that has any kind of an future.

Just trying to sense the mood - and find a way :-)

no. that apparently doesnt lay within their interest. it has been the subject, while there was some initiative to improve the voodoo drivers. also maybe karlos has something to tell about it, since he has been improving os4 w3d permedia drivers (under nda), i would be surprised if he was allowed to port his work to amiga (68k). and this politics while not exactly community friendly is perfectly within their rights and reason to force os4 upon amiga enthusiasts.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: nicholas on September 11, 2013, 08:56:31 PM
@John

Have you asked over at EAB or A1K yet? They seem to have most of the 68k coding wizards posting there now.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 11, 2013, 11:01:09 PM
dont fool yourself. you will hardly find anyone that writes a 3d driver on these sites. there are guys who can optimize 68k asm alright, but writing a whole 3d driver in asm from the scratch, i dont see it happen, moreover i see no sense in it. i think the scene is easy to oversee and i could count people who come into question on fingers of one hand. but none of them will actually do it. some sort of wrapper upon existing solution involving wazp3d, aros gallium or something like that is the best bet until some more really talented coders discover amiga, and i guess it will not happen with the current notions and attitudes.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: XDelusion on September 11, 2013, 11:13:59 PM
Really, they forced Ratte to stop?!?! That is SO FREAKING LAME!!!!!!!
Title: Re: Bounty for better Radeon Mediator drivers
Post by: Cosmos on September 12, 2013, 04:25:03 AM
Quote from: XDelusion;747679
Really, they forced Ratte to stop?!?! That is SO FREAKING LAME!!!!!!!


Yes, and it's a shame...
Title: Re: Bounty for better Radeon Mediator drivers
Post by: kickstart on September 12, 2013, 04:48:24 AM
Sometimes (many times) is like os4 is against the real amiga.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: XDelusion on September 12, 2013, 04:57:29 AM
What they should have done was put him on pay roll.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: XDelusion on September 12, 2013, 09:59:08 PM
So does anyone have those banned drivers handy for Amithlon? I'd REALLY like to get my hands on them. I'd rather have a better Amithlon over OS 4 any day at this point, especially considering the company politics... makes me not want to support them in the least.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: fishy_fiz on September 13, 2013, 02:03:45 AM
Yeah, I'd love to get my hands on ratte's openpci radeon drivers as well.
I still used, and enjoy amithlon as well, so had been watching the drivers progress since they were 1st announced. Unfortunately I didnt download them in time and they were removed before I had a chance to try them.

As for the thread, I'd be interested in a bounty if it comes to that, just as long as theyre openpci.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: Methuselas on September 13, 2013, 04:31:07 AM
I *THINK* I have them...... Not sure and I just moved, so my server's not setup yet. I'll check and see if I have them still. I haven't used Amithlon in a while.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: smf on September 13, 2013, 07:59:02 AM
better Radeon Mediator drivers for os3.x (warp3d support) would be nice, the lack of warp3d drivers for os3.1 is the main reason for me  to not use my radeon card and buy os4.1 classic.

The main os in my A4000 is and always will be os3.9 but it would be cool to play around with os4.1 on it too but but without better radeon drivers for 3.9 it wont happen because Voodoo3 is the best gfx card i can have on os3.9 and with os4.1 it's radeon. I would have to swap cards to switch OS or live with the slow voodoo3 when using 4.1 (i have tested voodoo3 on my pegasos2 and it's dead slow with some of the "modern" features in  os4 that i don't want to live without)
Title: Re: Bounty for better Radeon Mediator drivers
Post by: JohnFante on September 13, 2013, 09:00:11 AM
Quote from: Methuselas;747815
I *THINK* I have them...... Not sure and I just moved, so my server's not setup yet. I'll check and see if I have them still. I haven't used Amithlon in a while.

If you have them it would be great. :-)

Regarding the bounty I will send Elbox a mail and see how they react.

A first step would be to get better OS3.9 Radeon drivers. Warp3D/Wazp3D support for Radeon the next step.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: woof on September 13, 2013, 09:37:27 AM
Hello

You wrote
"for OS.39 suck They lack speed and there is no Warp3D support.
On OS4.1 Classic it is the other way around. The Voodoo drivers are slow and Warp3D support is not good"

Excuse my ignorance but why does the old  (good) 3.9 driver dont works on os4.1 ?
(After all they are both AmigaOS .library (ies) )
Is it because the 3.9 one is 68k and the 4.1 ppc ?
Or perhaps because 3.9 driver is Warp3D v4 and os4.1 is v5 ?
I mean does old 68k progs (that still use v4) works well on os4.1 with the old driver ? and only
new progs (that use v5) dont works  ?

>aros(68k), introduce pci bridges support to it, make gallium work on it, and introduce wazp3d to gallium wrapper. sounds complex and im not sure it would actually work, but thats what i would try. mind you i have pursued this subject few years ago and turned to aros as result anyway
I fully agree with Wawa :-)

The way is :
compile Mesa for Aros 68k : first as software
then make (hardware) gallium  works with  it on Aros 68k (pci stuff...)
Recompile Wazp3D-Aros (that already include an OpenGL renderer) on Aros 68k
 

Alain Thellier
Title: Re: Bounty for better Radeon Mediator drivers
Post by: JohnFante on September 13, 2013, 09:51:13 AM
Quote from: woof;747835
Hello

You wrote
"for OS.39 suck They lack speed and there is no Warp3D support.
On OS4.1 Classic it is the other way around. The Voodoo drivers are slow and Warp3D support is not good"

Excuse my ignorance but why does the old  (good) 3.9 driver dont works on os4.1 ?
(After all they are both AmigaOS .library (ies) )
Is it because the 3.9 one is 68k and the 4.1 ppc ?

In OS4.1 all graphicsdrivers etc. are PPC native. I have not tried to use the OS3.9 version of the Voodoo drivers. I will give that a go - but I do not have high expectations. I have used old 3.9 drivers with my Tocatta sound drivers but the result was not very effective (high CPU load).

Quote from: woof;747835
Or perhaps because 3.9 driver is Warp3D v4 and os4.1 is v5 ?
I mean does old 68k progs (that still use v4) works well on os4.1 with the old driver ? and only
new progs (that use v5) dont works  ?

I have not really tested that.

Quote from: woof;747835
The way is :
compile Mesa for Aros 68k : first as software
then make (hardware) gallium  works with  it on Aros 68k (pci stuff...)
Recompile Wazp3D-Aros (that already include an OpenGL renderer) on Aros 68k
 
Alain Thellier

Are you the man for that project Alain? :-)
Title: Re: Bounty for better Radeon Mediator drivers
Post by: nicholas on September 13, 2013, 10:29:32 AM
Quote from: XDelusion;747777
So does anyone have those banned drivers handy for Amithlon? I'd REALLY like to get my hands on them. I'd rather have a better Amithlon over OS 4 any day at this point, especially considering the company politics... makes me not want to support them in the least.


AFAIK Ratte's 68k driver is not Amithlon specific as it uses openpci.library. OpenPCI works on Mediators so in theory his driver should also work on them too.

I wonder if his driver is in archive.org or Google cache?
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 13, 2013, 10:33:34 AM
once it was possible to use 68k rtg drivers for os4 classic. i mean the elbox voodoo drivers. you had to disable the native drivers first in the kicklayout though. i dont remember exact procedure, but it was not recommended and i dont know if this still works under 4.1.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 13, 2013, 10:45:51 AM
hi alain!
Quote
The way is :
compile Mesa for Aros 68k : first as software
done, but not working correctly yet. its the same on other v1 platforms. have bothered deadwood about a fix without result;)
Quote
then make (hardware) gallium  works with  it on Aros 68k (pci stuff...)
partly done, pci half implemented, pending since long even though sonic has promised to look into it. apparently the interrupts are still missing.
Quote
Recompile Wazp3D-Aros (that already include an OpenGL renderer) on Aros 68k
it compiles for 68k. alas it doesnt work, crashes afair. may that be the reason why aros wazp3d  is much bigger than your genuine 68k compile?
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 13, 2013, 12:05:12 PM
would have to check what version of wazp is the aros commited one. the genuine wazp worked last time i checked on aros68k. must confirm though..
Title: Re: Bounty for better Radeon Mediator drivers
Post by: billt on September 13, 2013, 12:50:19 PM
Quote from: JohnFante;747597
I am considering setting up a bounty to collect enough money to get Elbox to develop and release better Radeon drivers with Warp3D support. And to get the drivers open sourced :-)


I'm not sure if getting the drivers open-sourced is likely. The Picasso96 stuff was an under-NDA-only sort of thing, so no P96 drivers have ever been open-sourced. I'm not sure about CyberGraphics drivers if they are all encumbered with NDAs sa well, or if they can be open-sourced.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: OlafS3 on September 13, 2013, 01:30:00 PM
Quote from: kickstart;747698
Sometimes (many times) is like os4 is against the real amiga.

they want to sell their licenses and avoid everything that would support users on other platforms. I can remember reading a thread where someone asked one of the Friedens to compile something for another platform (I think it was Warp3D for 68k) and the answer was I could but I do not, buy AmigaOS when you want to use it (i do not remember the exact phrases). That shows the attitude...

As Wawa said only chance is AROS 68k. BTW the same is partly the case on MorphOS team too. Both "closed-source" and "commercial" camps want to defend their advantages and are not interested in sharing code in areas where they think they are better than the others. It is their right of course but it would have been better in a common interest to work together. But that is not the case (and will propably never be).

Of course it is different outside the "core" and even in the teams there are exceptions that are very open (like Fab from MorphOS-Team) but I think every technology that is seen as advantage like MUI4, P96, Cybergraphics 5 and so that is copyrighted will not opened or at least cross-compiled.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: cgutjahr on September 13, 2013, 01:56:01 PM
Quote from: OlafS3;747859
they want to sell their licenses and avoid everything that would support users on other platforms.

Just to set the record straight: The "they" in "they stopped Ratte" referred to Elbox, not Hyperion. Hyperion has absolutely nothing to do with this.

Boring detail, I know. Let's bash Hyperion anyway.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: OlafS3 on September 13, 2013, 02:03:02 PM
Quote from: cgutjahr;747860
Just to set the record straight: The "they" in "they stopped Ratte" referred to Elbox, not Hyperion. Hyperion has absolutely nothing to do with this.

Boring detail, I know. Let's bash Hyperion anyway.

thanks for clarification... but most of the discussion was about Warp3D (that I was referring to). Ratte was stopped by Elbox (and not Hyperion).
Title: Re: Bounty for better Radeon Mediator drivers
Post by: nicholas on September 13, 2013, 02:09:58 PM
Quote from: cgutjahr;747860
Just to set the record straight: The "they" in "they stopped Ratte" referred to Elbox, not Hyperion. Hyperion has absolutely nothing to do with this.

Boring detail, I know. Let's bash Hyperion anyway.


The "they" in the discussion about Warp3D is obviously Hyperion, the conversation had moved on from Ratte's driver.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 13, 2013, 02:40:34 PM
this is not a question of blame but of a possible solution. asking elbox or hyperion to open their drivers or whatever software is a waste of time. the only solution for community driven efforts is open source. and open source efforts must be distributed to be any effective. none will do any substantial work on amiga anymore all by himself. therefore since the only open amiga platform is aros, all efforts should concentrate there. sorry about that.

the amiga (68k) platform is in unfortunate situation, that all c coders automatically moved to ng solutions since c is portable and platform independent language. whoever is left, is likely asm coder dedicated to 68k. therefore the only possibility for amiga 68k to keep up is to stay in sync with the c code developed for some of the ng platforms. neither os4 nor mos have much interest to share with the genuine amiga platform, so it leaves us again with only aros as alternative. go compute.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: cgutjahr on September 13, 2013, 02:50:45 PM
Quote from: OlafS3;747861
but most of the discussion was about Warp3D

You replied to Kickstart, who was talking about the party who stopped Ratte. Maybe I should have replied to him instead, to make things clearer.

Quote

discussion was about Warp3D

If you're still complaining about Hyperion not supporting you more than a decade after they made it perfectly clear they had zero interest to do so, something's badly wrong with the platform you're supporting. As in: "dead".

As for Hyperion not producing W3D drivers for OS3: If you really need somebody to give you a vague promise right now, post a useless status update in twelve months and insult you for being ungrateful dicks two years from now on - I can do that aswell, no need to wait for official support from Hyperion. We could also include some prepayment scheme to complete the experience.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: nicholas on September 13, 2013, 03:00:20 PM
Quote from: cgutjahr;747866
You replied to Kickstart, who was talking about the party who stopped Ratte. Maybe I should have replied to him instead, to make things clearer.


If you're still complaining about Hyperion not supporting you more than a decade after they made it perfectly clear they had zero interest to do so, something's badly wrong with the platform you're supporting. As in: "dead".

As for Hyperion not producing W3D drivers for OS3: If you really need somebody to give you a vague promise right now, post a useless status update in twelve months and insult you for being ungrateful dicks two years from now on - I can do that aswell, no need to wait for official support from Hyperion. We could also include some prepayment scheme to complete the experience.


:roflmao:
Title: Re: Bounty for better Radeon Mediator drivers
Post by: JohnFante on September 13, 2013, 03:03:07 PM
Ok, guys. Lets keep this cool. :-)

I asked - in the 1. post. - if there was any interest in a bounty for better Radeon (with Warp3D) drivers for OS3.9.

I can now conclude that there is very little interest, and that there is substantial problems in getting Elbox and Hyperion to play ball.

Instead people point to Aros for 68K as a possible solution for better Radeon support.

Lets focus the discussion on how this could be done.

wawrzon describes the status to be:

Compile Mesa for Aros 68k: first as software done, but not working correctly yet. Its the same on other v1 platforms. have bothered deadwood about a fix without result

Make (hardware) gallium works with it on Aros 68k (pci stuff...): Partly done, pci half implemented, pending since long even though sonic has promised to look into it. apparently the interrupts are still missing.

Recompile Wazp3D-Aros (that already include an OpenGL renderer) on Aros 68k: It compiles for 68k. Alas it doesnt work, crashes afair. may that be the reason why aros wazp3d is much bigger than your genuine 68k compile?

Is this correct?
Title: Re: Bounty for better Radeon Mediator drivers
Post by: woof on September 13, 2013, 03:04:33 PM
>Are you the man for that project Alain? :-)
Absolutely NO: I dont have classic hardware anymore so I wont be able to test such a low-lewel thing


>Recompile Wazp3D-Aros
>>for 68k. alas it doesnt work, crashes....
Curious the software rendering should works
Perhaps there is a problem with the #define in Wazp3D.h as Aros is supposed to be Intel ordered and to use Mesa/Gallium

Does Wazp3D-Prefs works ? What say the Wazp3D debugger before crashing ?

Alain Thellier (aka Wazp3D)
Title: Re: Bounty for better Radeon Mediator drivers
Post by: nicholas on September 13, 2013, 03:10:23 PM
Quote from: JohnFante;747868
Ok, guys. Lets keep this cool. :-)

I asked - in the 1. post. - if there was any interest in a bounty for better Radeon (with Warp3D) drivers for OS3.9.

I can now conclude that there is very little interest, and that there is substantial problems in getting Elbox and Hyperion to play ball.

Instead people point to Aros for 68K as a possible solution for better Radeon support.

Lets focus the discussion on how this could be done.

wawrzon describes the status to be:

Compile Mesa for Aros 68k: first as software done, but not working correctly yet. Its the same on other v1 platforms. have bothered deadwood about a fix without result

Make (hardware) gallium works with it on Aros 68k (pci stuff...): Partly done, pci half implemented, pending since long even though sonic has promised to look into it. apparently the interrupts are still missing.

Recompile Wazp3D-Aros (that already include an OpenGL renderer) on Aros 68k: It compiles for 68k. Alas it doesnt work, crashes afair. may that be the reason why aros wazp3d is much bigger than your genuine 68k compile?

Is this correct?


Yep.

However, I'm willing to host Ratte's driver on a server located in a country not subject to the laws that Elbox threatened him with if need be.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: OlafS3 on September 13, 2013, 04:00:58 PM
Quote from: cgutjahr;747866
You replied to Kickstart, who was talking about the party who stopped Ratte. Maybe I should have replied to him instead, to make things clearer.


If you're still complaining about Hyperion not supporting you more than a decade after they made it perfectly clear they had zero interest to do so, something's badly wrong with the platform you're supporting. As in: "dead".

As for Hyperion not producing W3D drivers for OS3: If you really need somebody to give you a vague promise right now, post a useless status update in twelve months and insult you for being ungrateful dicks two years from now on - I can do that aswell, no need to wait for official support from Hyperion. We could also include some prepayment scheme to complete the experience.

"As in: "dead""

thanks for the tip. Now i finally know what is wrong with it :-)
Title: Re: Bounty for better Radeon Mediator drivers
Post by: OlafS3 on September 13, 2013, 04:03:23 PM
I think Wazp3D works on Aros 68k. I have installed it together this with Sormmesa2010 and run the old "Stormmesa-Demos".
Title: Re: Bounty for better Radeon Mediator drivers
Post by: nicholas on September 13, 2013, 04:34:23 PM
Quote from: OlafS3;747878
I think Wazp3D works on Aros 68k. I have installed it together this with Sormmesa2010 and run the old "Stormmesa-Demos".


Yes but it is software only rendering, the AROS specific version is hardware accelerated so that is the best bet to get working.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: OlafS3 on September 13, 2013, 04:44:36 PM
Quote from: nicholas;747881
Yes but it is software only rendering, the AROS specific version is hardware accelerated so that is the best bet to get working.

Hardware accellerated rendering would be of course better. But would it possible to adapt Gallium to use Wazp3D as layer? This solution would then run everywhere and we only need a adapted version of Wazp3D for f.e. a certain graphiccard.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 13, 2013, 04:45:22 PM
Quote from: OlafS3;747878
I think Wazp3D works on Aros 68k. I have installed it together this with Sormmesa2010 and run the old "Stormmesa-Demos".

i was speaking of wazp3d from aros contributions. alains pacakge from aminet worked.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 13, 2013, 05:16:01 PM
Quote from: JohnFante;747868
Ok, guys. Lets keep this cool. :-)
sure ;)
Quote
Is this correct?
the status is as you see that a considerable number of things is in place for this approach as you see.
the problem is it is impossible to get aros maintainers to take care of these as they have other priorities and the main 68k maintainer, jason, is busy at his day job. i have already problems to motivate anyone to fix smaller things. this is the reality of the situation alas.

the solution would be if another dediacted amiga (68k) coder with good c experience and some time and dedication filled into the gap even if led by hand by experienced aros devs. it is very unlikely to find anyone like that though.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 13, 2013, 05:40:00 PM
Quote from: woof;747870

Does Wazp3D-Prefs works ? What say the Wazp3D debugger before crashing ?

Alain Thellier (aka Wazp3D)

yes they work, what options?
Title: Re: Bounty for better Radeon Mediator drivers
Post by: woof on September 13, 2013, 06:28:42 PM
>What options

Renderer: hard
Enable debugger
Debug var
Debug val
Debug function
Debug  soft3d
Debuug sc & hard
Title: Re: Bounty for better Radeon Mediator drivers
Post by: Iggy on September 13, 2013, 06:32:42 PM
"I you build it, they will come"

Hey if you guys ask for this enough, someone is likely to pick it up and run with it.
That's just Amiga history.
'Dead' is never an option for us.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 13, 2013, 07:27:26 PM
Quote from: woof;747895
>What options

Renderer: hard
Enable debugger
Debug var
Debug val
Debug function
Debug  soft3d
Debuug sc & hard

it cant open gallium.hidd and gallium.library. which is selfexplanatory since i have not
loadresource gallium.hidd
but as is said gallium (softpipe) is not yet working correctly as such so i think we are starting from the wrong end? shouldnt i choose the wazp soft renderer for starters?

edit: how about soft to bitmap and soft to image options? also how do i log the output to a file? im getting immediate crashes in most cases.

btw you dont need amiga hardware (at least at this stage of debugging) you can run aros68k under uae of your choice. if you decompress aros nightly iso (along with the contrib archive to the same location) and boot from there (you can use genuine amiga 3.x roms if you cant use aros ones for some reason, aros will softkick itself upon the genuine kickstart).
Title: Re: Bounty for better Radeon Mediator drivers
Post by: woof on September 13, 2013, 07:57:46 PM
Yes if gallium softpipe is not ready
then u can try soft to image

It will just allow to know if wazp aros 68k is correct
But if gallium is not ready. ..  No much interesting

Alain
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 13, 2013, 09:06:54 PM
as i said i asked deadwood, but gallium softpipe on v1 is not working neither on 68k nor on x86. that means it runs, ocassionaly displays some bits here and there, but mostly displays only an empty window. so except the console output about fps and the like one would think it has crashed. so far gallium works only on aros v0 but since a number of things changed since then one would have to know what changed to fix it.

btw i tested soft to image and it crashed hard as well, but since i dont know how to catch the log to file i cant post the log.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: woof on September 13, 2013, 09:44:30 PM
>Catch the log
Are you kidding?
just do a redirection something like
 yourprog > ram:debug.txt

Alain
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 13, 2013, 10:36:56 PM
sorry ive forgot, didnt use that for long.
now, im not sure there is much to report. say, putting up stack and running warptest i choose 16bit screen as only available and get a black screen (turns dirty yellow after a click) thats all..
Code: [Select]
W3D_CheckDriver
 Wazp3D->drivertype=4
Hardware driver available
W3D_GetDrivers
Available drivers:
Wazp3D - Alain Thellier - Paris France 2012 - Beta 52
Supports format 0x7FFF
Hardware Driver
Primitives supported:
W3D_QueryDriver
 destfmt=2
 query=1
 W3D_Q_DRAW_POINT
 support=W3D_FULLY_SUPPORTED
[POINT] W3D_QueryDriver
 destfmt=2
 query=2
 W3D_Q_DRAW_LINE
 support=W3D_FULLY_SUPPORTED
[LINE] W3D_QueryDriver
 destfmt=2
 query=3
 W3D_Q_DRAW_TRIANGLE
 support=W3D_FULLY_SUPPORTED
[TRIANGLE]
Filtering:
W3D_QueryDriver
 destfmt=2
 query=13
 W3D_Q_BILINEARFILTER
 support=W3D_FULLY_SUPPORTED
[BI-FILTER] W3D_QueryDriver
 destfmt=2
 query=14
 W3D_Q_MMFILTER
 support=W3D_FULLY_SUPPORTED
[MM-FILTER]


W3D_RequestMode
 W3D_SMR_TYPE,4,
 W3D_SMR_SIZEFILTER,1,
 W3D_SMR_DESTFMT,-2,
 ASLSM_MinWidth,640,
 ASLSM_MaxWidth,641,
 ASLSM_MinHeight,480,
 ASLSM_MaxHeight,481,
ScreenModeFilter[69632]320 X 200 X 8
ScreenModeFilter[102400]640 X 200 X 8
ScreenModeFilter[102432]1280 X 200 X 8
ScreenModeFilter[135168]320 X 256 X 8
ScreenModeFilter[167936]640 X 256 X 8
ScreenModeFilter[167968]1280 X 256 X 8
ScreenModeFilter[69636]320 X 400 X 8
ScreenModeFilter[102404]640 X 400 X 8
ScreenModeFilter[102436]1280 X 400 X 8
ScreenModeFilter[135172]320 X 512 X 8
ScreenModeFilter[167940]640 X 512 X 8
ScreenModeFilter[167972]1280 X 512 X 8
ScreenModeFilter[268439808]320 X 200 X 8
ScreenModeFilter[268442112]320 X 200 X 16
ScreenModeFilter[268505344]320 X 240 X 8
ScreenModeFilter[268507648]320 X 240 X 16
ScreenModeFilter[269029632]320 X 256 X 8
ScreenModeFilter[269031936]320 X 256 X 16
ScreenModeFilter[268570880]640 X 400 X 8
ScreenModeFilter[268573184]640 X 400 X 16
ScreenModeFilter[268636416]640 X 480 X 8
ScreenModeFilter[268638720]640 X 480 X 16
ScreenModeFilter[269095168]640 X 512 X 8
ScreenModeFilter[269097472]640 X 512 X 16
ScreenModeFilter[277025024]720 X 480 X 8
ScreenModeFilter[277027328]720 X 480 X 16
ScreenModeFilter[277090560]720 X 576 X 8
ScreenModeFilter[277092864]720 X 576 X 16
ScreenModeFilter[268701952]800 X 600 X 8
ScreenModeFilter[268704256]800 X 600 X 16
ScreenModeFilter[268767488]1024 X 768 X 8
ScreenModeFilter[268769792]1024 X 768 X 16
ScreenModeFilter[268833024]1152 X 864 X 8
ScreenModeFilter[268835328]1152 X 864 X 16
ScreenModeFilter[277745920]1280 X 720 X 8
ScreenModeFilter[277748224]1280 X 720 X 16
ScreenModeFilter[277811456]1280 X 768 X 8
ScreenModeFilter[277813760]1280 X 768 X 16
ScreenModeFilter[277876992]1280 X 800 X 8
ScreenModeFilter[277879296]1280 X 800 X 16
ScreenModeFilter[278008064]1280 X 960 X 8
ScreenModeFilter[278010368]1280 X 960 X 16
ScreenModeFilter[268898560]1280 X 1024 X 8
ScreenModeFilter[268900864]1280 X 1024 X 16
ScreenModeFilter[278991104]1360 X 768 X 8
ScreenModeFilter[278993408]1360 X 768 X 16
ScreenModeFilter[279580928]1600 X 900 X 8
ScreenModeFilter[279583232]1600 X 900 X 16
ScreenModeFilter[285151488]1600 X 1024 X 8
ScreenModeFilter[285153792]1600 X 1024 X 16
ScreenModeFilter[278270208]1600 X 1200 X 8
ScreenModeFilter[278272512]1600 X 1200 X 16
ScreenModeFilter[278335744]1680 X 1050 X 8
ScreenModeFilter[278338048]1680 X 1050 X 16
ScreenModeFilter[278401280]1920 X 1080 X 8
ScreenModeFilter[278403584]1920 X 1080 X 16
ScreenModeFilter[278466816]1920 X 1200 X 8
ScreenModeFilter[278469120]1920 X 1200 X 16
 ModeID=268638720
W3D_CreateContext
 error=146790520
 taglist=146790380
 tag=-2145386488
 data=268638720
 W3D_CC_MODEID,268638720,
 tag=-2145386496
 data=139786000
 W3D_CC_BITMAP,139786000,
 tag=-2145386495
 data=0
 W3D_CC_YOFFSET,0,
 tag=-2145386494
 data=2
 W3D_CC_DRIVERTYPE,2,
 tag=-2145386490
 data=0
 W3D_CC_DOUBLEHEIGHT,0,
 tag=-2145386489
 data=1
 W3D_CC_FAST,1,
 context->drivertype=W3D_DRIVER_BEST
 WC->window->LeftEdge=0
 WC->window->TopEdge=0
 WC->window->Width=640
 WC->window->Height=480
 sizeof(struct WAZP3D_context)=85990
 sizeof(struct WAZP3D_texture)=252
 sizeof(struct WAZP3D_blendstage)=18
 sizeof(struct state3D)=56
SetDrawRegion
setting the bitmap
 StateON(W3D_DOUBLEHEIGHT)=0
 yoffset=0
 context->width=640
 context->height=480
 WC->large=0
 WC->high=0
 SameSize=0
 WC->large=640
 WC->high=480
no scissor yet
 context->scissor.left=0
 context->scissor.top=0
 context->scissor.width=640
 context->scissor.height=480
 sizeof(union pixel3D)=106
 sizeof(union oper3D )=4
 sizeof(struct fragbuffer3D)=24
 sizeof(struct HARD3D_context)=68246
 sizeof(struct SOFT3D_context)=2525636
 sizeof(struct SOFT3D_mipmap)=3072
 sizeof(struct SOFT3D_texture)=36936
 sizeof(struct state3D)=56
SOFT3D_Start
 PrefsWazp3D=140103048
SOFT3D_Start:OK
Bitmap size changed !!!!
SOFT3D_AllocImageBuffer
 large=640
 high=480
SetImage
 Image8=149316592
 x=0
 y=0
 large=640
 high=480
 bits=32
 offset=0
 SC->Image8=149316592
SOFT3D_SetBitmap
 bm=139786000
 bmdata=270330368
 bmformat=5
SetImage to ImageBuffer32
SetImage
 Image8=149316592
 x=0
 y=0
 large=640
 high=480
 bits=32
 offset=0
 SC->Image8=149316592
SOFT3D_SetClipping
 xmin=0
 xmax=639
 ymin=0
 ymax=479
 context->width=640
 context->height=480
 context->depth=16
 context->bprow=80
 context->format=0
 context->drawregion->BytesPerRow=80
 context->drawregion->Rows=480
 context->drawregion->Flags=146
 context->drawregion->Depth=8
 context->drawregion->pad=24904
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 13, 2013, 11:41:49 PM
cow3d rendering to bitmap shows an empty window till clicked into the window bar and moved around the screen. then i can recognize cow vertices as if with some texture projected frontally over them for a moment and i get a system crash and crash loop. uae log gives me some memory (alloc?) errors but hard to catch due to reset loop. here the end of the w3d log, but i think it doesnt say much since the moment of the crash is random:

Code: [Select]

...
PixelsTex32ToBuffer
Pass 7:FunctionOut
 (void *)PixelsOut16
Flush done...
SOFT3D_SetDrawFunctions
 FlatRGBA: 102 117 95 255
Gouraud to Flat
 (void *)Ztest_zless_update
CurrentRGBA dont have alpha
Functions set with TexEnvMode1 BlendMode33
ChangeSof±Point
 SC->state.PointSize=1
DrawPrimitive
ZbufferCheck
SOFT3D_DrawPrimitive Pnb 3 (0)
 ST=151374488
 Texture2_256X256X32.RAW
 W3D_PRIMITIVE_TRIANGLES
DrawPolyP Pnb 3  
FlatRGBA changed
ColorChange
ColorChange
DrawPolyPix
 SC->PolyPnb=3
Edge_Tex
Edge_Tex
Edge_Tex
Poly_Persp0_Tex
 maxlarge=28
Fill_Tex
-----------
SetTexStates
 WC->state:Changed0 UseTex1 ST151374488 ZMode9 TexEnvMode1 BlendMode33 UseGouraud1 UseFog0
 texture 139954874 was 139954874
Adjusting WC->state ....
 WC->state.BlendMode=33
 WC->DstFunc=15
 WC->SrcFunc=15
 WC->state.BlendMode=33
 WC->state:Changed1 UseTex1 ST151374488 ZMode9 TexEnvMode1 BlendMode33 UseGouraud1 UseFog0
SOFT3D_SetDrawState
 state:Changed1 UseTex1 ST151374488 ZMode9 TexEnvMode1 BlendMode33 UseGouraud1 UseFog0
 CurrentRGBA: 255 255 255 255
SOFT3D_Flush
 SC->FragSize2=4
Pass3: FunctionBlend (transparency)
PixelsTex32ToBuffer
Pass 7:FunctionOut
 (void *)PixelsOut16
Flush done...
SOFT3D_SetDrawFunctions
 FlatRGBA: 90 117 87 255
Gouraud to Flat
 (void *)Ztest_zless_update
CurrentRGBA dont have alpha
Functions set with TexEnvMode1 BlendMode33
ChangeSof±Point
 SC->state.PointSize=1
DrawPrimitive
ZbufferCheck
SOFT3D_DrawPrimitive Pnb 3 (0)
 ST=151374488
 Texture2_256X256X32.RAW
 W3D_PRIMITIVE_TRIANGLES
DrawPolyP Pnb 3  
FlatRGBA changed
ColorChange
ColorChange
DrawPolyPix
 SC->PolyPnb=3
Edge_Tex
Edge_Tex
Edge_Tex
Poly_Persp0_Tex
 maxlarge=27
Fill_Tex
-----------
SetTexStates
 WC->state:Changed0 UseTex1 ST151374488 ZMode9 TexEnvMode1 BlendMode33 UseGouraud1 UseFog0
 texture 139954874 was 139954874
Adjusting WC->state ....
 WC->state.BlendMode=33
 WC->DstFunc=15
 WC->SrcFunc=15
 WC->state.BlendMode=33
 WC->state:Changed1 UseTex1 ST151374488 ZMode9 TexEnvMode1 BlendMode33 UseGouraud1 UseFog0
SOFT3D_SetDrawState
 state:Changed1 UseTex1 ST151374488 ZMode9 TexEnvMode1 BlendMode33 UseGouraud1 UseFog0
 CurrentRGBA: 255 255 255 255
SOFT3D_Flush
 SC->FragSize2=28
Pass3: FunctionBlend (transparency)
PixelsTex32ToBuffer
Pass 7:FunctionOut
 (void *)PixelsOut16
Flush done...
SOFT3D_SetDrawFunctions
 FlatRGBA: 97 119 95 255
Gouraud to Flat
 (void *)Ztest_zless_update
CurrentRGBA dont have alpha
Functions set with TexEnvMode1 BlendMode33
ChangeSof±Point
 SC->state.PointSize=1
DrawPrimitive
ZbufferCheck
SOFT3D_DrawPrimitive Pnb 3 (0)
 ST=151374488
 Texture2_256X256X32.RAW
 W3D_PRIMITIVE_TRIANGLES
DrawPolyP Pnb 3  
FlatRGBA changed
ColorChange
ColorChange
DrawPolyPix
 SC->PolyPnb=3
Edge_Tex
Edge_Tex
Poly_Persp0_Tex
 maxlarge=2
Fill_Tex
-----------
SetTexStates
 WC->state:Changed0 UseTex1 ST151374488 ZMode9 TexEnvMode1 BlendMode33 UseGouraud1 UseFog0
 texture 139954874 was 139954874
Adjusting WC->state ....
 WC->state.BlendMode=33
 WC->DstFunc=15
 WC->SrcFunc=15
 WC->state.BlendMode=33
 WC->state:Changed1 UseTex1 ST151374488 ZMode9 TexEnvMode1 BlendMode33 UseGouraud1 UseFog0
SOFT3D_SetDrawState
 state:Changed1 UseTex1 ST151374488 ZMode9 TexEnvMode1 BlendMode33 UseGouraud1 UseFog0
 CurrentRGBA: 255 255 255 255
SOFT3D_Flush
 SC->FragSize2=3
Pass3: FunctionBlend (transparency)
PixelsTex32ToBuffer
Pass 7:FunctionOut
 (void *)PixelsOut16
Flush done...
SOFT3D_SetDrawFunctions
 FlatRGBA: 92 36 107 255
Gouraud to Flat
 (void *)Ztest_zless_update
CurrentRGBA dont have alpha
Functions set with TexEnvMode1 BlendMode33
ChangeSof±Point
 SC->state.PointSize=1
DrawPrimitive
ZbufferCheck
SOFT3D_DrawPrimitive Pnb 3 (0)
 ST=151374488
 Texture2_256X256X32.RAW
 W3D_PRIMITIVE_TRIANGLES
DrawPolyP Pnb 3  
FlatRGBA changed
ColorChange
ColorChange
DrawPolyPix
 SC->PolyPnb=3
Edge_Tex
Edge_Tex
Poly_Persp0_Tex
 maxlarge=6
Fill_Tex
-----------
SetTexStates
 WC->state:Changed0 UseTex1 ST151374488 ZMode9 TexEnvMode1 BlendMode33 UseGouraud1 UseFog0
 texture 139954874 was 139954874
Adjusting WC->state ....
 WC->state.BlendMode=33
 WC->DstFunc=15
 WC->SrcFunc=15
 WC->state.BlendMode=33
 WC->state:Changed1 UseTex1 ST151374488 ZMode9 TexEnvMode1 BlendMode33 UseGouraud1 UseFog0
SOFT3D_SetDrawState
 state:Changed1 UseTex1 ST151374488 ZMode9 TexEnvMode1 BlendMode33 UseGouraud1 UseFog0
 CurrentRGBA: 255 255 255 255
SOFT3D_Flush
 SC->FragSize2=7
Pass3: FunctionBlend (transparency)
PixelsTex32ToBuffer
Pass 7:FunctionOut
 (void *)PixelsOut16
Flush done...
SOFT3D_SetDrawFunctions
 FlatRGBA: 90 35 113 255
Gouraud to Flat
 (void *)Ztest_zless_update
CurrentRGBA dont have alpha
Functions set with TexEnvMode1 BlendMode33
ChangeSof±Point
 SC->state.PointSize=1
DrawPrimitive
ZbufferCheck
SOFT3D_DrawPrimitive Pnb 3 (0)
 ST=151374488
 Texture2_256X256X32.RAW
 W3D_PRIMITIVE_TRIANGLES
DrawPolyP Pnb 3  
FlatRGBA changed
ColorChange
ColorChange
DrawPolyPix
 SC->PolyPnb=3
Edge_Tex
Edge_Tex
Poly_Persp0_Tex
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 13, 2013, 11:48:52 PM
now for the fun of it i turned on gallium softpipe nevertheless. the errors in winuaelog are the same running cow3d in hardware (but i guess it will not tell you much):

Code: [Select]

Illegal bput at 0ccb38e8=00000000 PC=8784e8c
Illegal bput at 0ccc38e8=00000004 PC=8784ea8
Illegal bput at 0ccd38e8=00000002 PC=8784eb6
Illegal bput at 0dcb38e8=00000000 PC=8784e86
Illegal bput at 0dcc38e8=00000005 PC=8784e86
Illegal bput at 0dcd38e8=00000002 PC=8784e86
Illegal bput at 0ecb38e8=00000000 PC=8784e86
Illegal bput at 0ecc38e8=00000006 PC=8784e86
Illegal bput at 0ecd38e8=00000003 PC=8784e86
Illegal bput at 0fcb38e8=00000000 PC=8784e86
Illegal bput at 0fcc38e8=00000007 PC=8784e86
Illegal bput at 0fcd38e8=00000003 PC=8784e86
Illegal bput at 14cb38e8=00000000 PC=8784e86
Illegal bput at 14cc38e8=0000000c PC=8784e86
Illegal bput at 14cd38e8=00000006 PC=8784e86
Illegal bput at 15cb38e8=00000000 PC=8784e86
Illegal bput at 15cc38e8=0000000d PC=8784e86
Illegal bput at 15cd38e8=00000006 PC=8784e86
Illegal bput at 16cb38e8=00000000 PC=8784e86
Illegal bput at 16cc38e8=0000000e PC=8784e86
Illegal bput at 16cd38e8=00000007 PC=8784e86
Illegal bput at 17cb38e8=00000000 PC=8784e86
Illegal bput at 17cc38e8=0000000f PC=8784e86
Illegal bput at 17cd38e8=00000007 PC=8784e86
Illegal bput at 18cb38e8=00000000 PC=8784e86
Illegal bput at 18cc38e8=00000010 PC=8784e86
Illegal bput at 18cd38e8=00000008 PC=8784e86
Illegal bput at 19cb38e8=00000000 PC=8784e86
Illegal bput at 19cc38e8=00000011 PC=8784e86
Illegal bput at 19cd38e8=00000008 PC=8784e86
Illegal bput at 1acb38e8=00000000 PC=8784e86
Illegal bput at 1acc38e8=00000012 PC=8784e86
Illegal bput at 1acd38e8=00000009 PC=8784e86
Illegal bput at 1bcb38e8=00000000 PC=8784e86
Illegal bput at 1bcc38e8=00000013 PC=8784e86
Illegal bput at 1bcd38e8=00000009 PC=8784e86
Illegal bput at 1ccb38e8=00000000 PC=8784e86
Illegal bput at 1ccc38e8=00000014 PC=8784e86
Illegal bput at 1ccd38e8=0000000a PC=8784e86
Illegal bput at 1dcb38e8=00000000 PC=8784e86
Illegal bput at 1dcc38e8=00000015 PC=8784e86
Illegal bput at 1dcd38e8=0000000a PC=8784e86
Illegal bput at 1ecb38e8=00000000 PC=8784e86
Illegal bput at 1ecc38e8=00000016 PC=8784e86
Illegal bput at 1ecd38e8=0000000b PC=8784e86
Illegal bput at 1fcb38e8=00000000 PC=8784e86
Illegal bput at 1fcc38e8=00000017 PC=8784e86
Illegal bput at 1fcd38e8=0000000b PC=8784e86
Illegal bput at 20cb38e8=00000000 PC=8784e86
Illegal bput at 20cc38e8=00000018 PC=8784e86
Illegal bput at 20cd38e8=0000000c PC=8784e86
Illegal bput at 21cb38e8=00000000 PC=8784e86
Illegal bput at 21cc38e8=00000019 PC=8784e86
Illegal bput at 21cd38e8=0000000c PC=8784e86
Illegal bput at 22cb38e8=00000000 PC=8784e86
Illegal bput at 22cc38e8=0000001a PC=8784e86
Illegal bput at 22cd38e8=0000000d PC=8784e86
Illegal bput at 23cb38e8=00000000 PC=8784e86
Illegal bput at 23cc38e8=0000001b PC=8784e86
Illegal bput at 23cd38e8=0000000d PC=8784e86
Illegal bput at 24cb38e8=00000000 PC=8784e86
Illegal bput at 24cc38e8=0000001c PC=8784e86
Illegal bput at 24cd38e8=0000000e PC=8784e86
Illegal bput at 25cb38e8=00000000 PC=8784e86
Illegal bput at 25cc38e8=0000001d PC=8784e86
Illegal bput at 25cd38e8=0000000e PC=8784e86
Illegal bput at 26cb38e8=00000000 PC=8784e86
Illegal bput at 26cc38e8=0000001e PC=8784e86
Illegal bput at 26cd38e8=0000000f PC=8784e86
Illegal bput at 27cb38e8=00000000 PC=8784e86
Illegal bput at 27cc38e8=0000001f PC=8784e86
Illegal bput at 27cd38e8=0000000f PC=8784e86
Illegal bput at 28cb38e8=00000000 PC=8784e86
Illegal bput at 28cc38e8=00000020 PC=8784e86
Illegal bput at 28cd38e8=00000010 PC=8784e86
Illegal bput at 29cb38e8=00000000 PC=8784e86
Illegal bput at 29cc38e8=00000021 PC=8784e86
Illegal bput at 29cd38e8=00000010 PC=8784e86
Illegal bput at 2acb38e8=00000000 PC=8784e86
Illegal bput at 2acc38e8=00000022 PC=8784e86
Illegal bput at 2acd38e8=00000011 PC=8784e86
Illegal bput at 2bcb38e8=00000000 PC=8784e86
Illegal bput at 2bcc38e8=00000023 PC=8784e86
Illegal bput at 2bcd38e8=00000011 PC=8784e86
Illegal bput at 2ccb38e8=00000000 PC=8784e86
Illegal bput at 2ccc38e8=00000024 PC=8784e86
Illegal bput at 2ccd38e8=00000012 PC=8784e86
Illegal bput at 2dcb38e8=00000000 PC=8784e86
Illegal bput at 2dcc38e8=00000025 PC=8784e86
Illegal bput at 2dcd38e8=00000012 PC=8784e86
Illegal bput at 2ecb38e8=00000000 PC=8784e86
Illegal bput at 2ecc38e8=00000026 PC=8784e86
Illegal bput at 2ecd38e8=00000013 PC=8784e86
Illegal bput at 2fcb38e8=00000000 PC=8784e86
Illegal bput at 2fcc38e8=00000027 PC=8784e86
Illegal bput at 2fcd38e8=00000013 PC=8784e86
Illegal bput at 30cb38e8=00000000 PC=8784e86
Illegal bput at 30cc38e8=00000028 PC=8784e86
Illegal bput at 30cd38e8=00000014 PC=8784e86
Illegal bput at 31cb38e8=00000000 PC=8784e86
Illegal bput at 31cc38e8=00000029 PC=8784e86
Illegal bput at 31cd38e8=00000014 PC=8784e86
Illegal bput at 32cb38e8=00000000 PC=8784e86
Illegal bput at 32cc38e8=0000002a PC=8784e86
Illegal bput at 32cd38e8=00000015 PC=8784e86
Illegal bput at 33cb38e8=00000000 PC=8784e86
Illegal bput at 33cc38e8=0000002b PC=8784e86
Illegal bput at 33cd38e8=00000015 PC=8784e86
Illegal bput at 34cb38e8=00000000 PC=8784e86
Illegal bput at 34cc38e8=0000002c PC=8784ea4
Illegal bput at 34cd38e8=00000016 PC=8784ea4
Illegal bput at 35cb38e8=00000000 PC=8784e86
Illegal bput at 35cc38e8=0000002d PC=8784ea4
Illegal bput at 35cd38e8=00000016 PC=8784ea4
Illegal bput at 36cb38e8=00000000 PC=8784e86
Illegal bput at 36cc38e8=0000002e PC=8784ea4
Illegal bput at 36cd38e8=00000017 PC=8784ea4
Illegal bput at 37cb38e8=00000000 PC=8784e86
Illegal bput at 37cc38e8=0000002f PC=8784ea4
Illegal bput at 37cd38e8=00000017 PC=8784ea4
Illegal bput at 38cb38e8=00000000 PC=8784e86
Illegal bput at 38cc38e8=00000030 PC=8784ea4
Illegal bput at 38cd38e8=00000018 PC=8784ea4
Illegal bput at 39cb38e8=00000000 PC=8784e86
Illegal bput at 39cc38e8=00000031 PC=8784ea4
Illegal bput at 39cd38e8=00000018 PC=8784ea4
Illegal bput at 3acb38e8=00000000 PC=8784e86
Illegal bput at 3acc38e8=00000032 PC=8784ea4
Illegal bput at 3acd38e8=00000019 PC=8784ea4
Illegal bput at 3bcb38e8=00000000 PC=8784e86
Illegal bput at 3bcc38e8=00000033 PC=8784ea4
Illegal bput at 3bcd38e8=00000019 PC=8784ea4
Illegal bput at 3ccb38e8=00000000 PC=8784e86
Illegal bput at 3ccc38e8=00000034 PC=8784ea4
Illegal bput at 3ccd38e8=0000001a PC=8784ea4
Illegal bput at 3dcb38e8=00000000 PC=8784e86
Illegal bput at 3dcc38e8=00000035 PC=8784ea4
Illegal bput at 3dcd38e8=0000001a PC=8784ea4
Illegal bput at 3ecb38e8=00000000 PC=8784e86
Illegal bput at 3ecc38e8=00000036 PC=8784ea4
Illegal bput at 3ecd38e8=0000001b PC=8784ea4
Illegal bput at 3fcb38e8=00000000 PC=8784e86
Illegal bput at 3fcc38e8=00000037 PC=8784ea4
Illegal bput at 3fcd38e8=0000001b PC=8784ea4
Illegal bput at 40cb38e8=00000000 PC=8784e86
Illegal bput at 40cc38e8=00000038 PC=8784ea4
Illegal bput at 40cd38e8=0000001c PC=8784ea4
Illegal bput at 41cb38e8=00000000 PC=8784e86
Illegal bput at 41cc38e8=00000039 PC=8784ea4
Illegal bput at 41cd38e8=0000001c PC=8784ea4
Illegal bput at 42cb38e8=00000000 PC=8784e86
Illegal bput at 42cc38e8=0000003a PC=8784ea4
Illegal bput at 42cd38e8=0000001d PC=8784ea4
Illegal bput at 43cb38e8=00000000 PC=8784e86
Illegal bput at 43cc38e8=0000003b PC=8784ea4
Illegal bput at 43cd38e8=0000001d PC=8784ea4
Illegal bput at 44cb38e8=00000000 PC=8784e86
Illegal bput at 44cc38e8=0000003c PC=8784ea4
Illegal bput at 44cd38e8=0000001e PC=8784ea4
Illegal bput at 45cb38e8=00000000 PC=8784e86
Illegal bput at 45cc38e8=0000003d PC=8784ea4
Illegal bput at 45cd38e8=0000001e PC=8784ea4
Illegal bput at 46cb38e8=00000000 PC=8784e86
Illegal bput at 46cc38e8=0000003e PC=8784ea4
Illegal bput at 46cd38e8=0000001f PC=8784ea4
Illegal bput at 47cb38e8=00000000 PC=8784e86
Illegal bput at 47cc38e8=0000003f PC=8784ea4
Illegal bput at 47cd38e8=0000001f PC=8784ea4
Illegal bput at 48cb38e8=00000000 PC=8784e86
Illegal bput at 48cc38e8=00000040 PC=8784ea4
Illegal bput at 48cd38e8=00000020 PC=8784ea4
Illegal kickmem bput at 00e038e8
Illegal kickmem bput at 00e038e8
Illegal kickmem bput at 00e138e8
Illegal kickmem bput at 00e038e8
Illegal kickmem bput at 00e138e8
Illegal kickmem bput at 00e238e8
Illegal kickmem bput at 00e138e8
Illegal kickmem bput at 00e238e8
Illegal kickmem bput at 00e338e8
Illegal kickmem bput at 00e238e8
Illegal kickmem bput at 00e338e8
Illegal kickmem bput at 00e438e8
Illegal kickmem bput at 00e338e8
Illegal kickmem bput at 00e438e8
Illegal kickmem bput at 00e538e8
Illegal kickmem bput at 00e438e8
Illegal kickmem bput at 00e538e8
Illegal kickmem bput at 00e638e8
Illegal kickmem bput at 00e538e8
Illegal kickmem bput at 00e638e8
Illegal kickmem bput at 00e738e8
Illegal kickmem bput at 00e638e8
Illegal kickmem bput at 00e738e8
Illegal kickmem bput at 00e738e8
Illegal kickmem bput at 00f838e8
Illegal kickmem bput at 00f838e8
Illegal kickmem bput at 00f938e8
Illegal kickmem bput at 00f838e8
Illegal kickmem bput at 00f938e8
Illegal kickmem bput at 00fa38e8
Illegal kickmem bput at 00f938e8
Illegal kickmem bput at 00fa38e8
Illegal kickmem bput at 00fb38e8
Illegal kickmem bput at 00fa38e8
Illegal kickmem bput at 00fb38e8
Illegal kickmem bput at 00fc38e8
Illegal kickmem bput at 00fb38e8
Illegal kickmem bput at 00fc38e8
Illegal kickmem bput at 00fd38e8
Illegal kickmem bput at 00fc38e8
Illegal kickmem bput at 00fd38e8
Illegal kickmem bput at 00fe38e8
Illegal kickmem bput at 00fd38e8
Illegal kickmem bput at 00fe38e8
Illegal kickmem bput at 00ff38e8
Illegal kickmem bput at 00fe38e8
Illegal kickmem bput at 00ff38e8
Illegal kickmem bput at 00ff38e8
Illegal instruction: 0009 at 0BF638E8 -> 0BF256C0
A-Trap ae2f at 87938e8 (283E38E8)
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 13, 2013, 11:50:57 PM
while here is what w3d log tells us till the crash:

Code: [Select]
error=140532066
 taglist=147379944
 tag=-2145386488
 data=268704256
 W3D_CC_MODEID,268704256,
 tag=-2145386496
 data=139872440
 W3D_CC_BITMAP,139872440,
 tag=-2145386495
 data=0
 W3D_CC_YOFFSET,0,
 tag=-2145386494
 data=2
 W3D_CC_DRIVERTYPE,2,
 tag=-2145386490
 data=0
 W3D_CC_DOUBLEHEIGHT,0,
 tag=-2145386489
 data=1
 W3D_CC_FAST,1,
 context->drivertype=W3D_DRIVER_BEST
 WC->window->LeftEdge=80
 WC->window->TopEdge=60
 WC->window->Width=648
 WC->window->Height=499
 sizeof(struct WAZP3D_context)=85990
 sizeof(struct WAZP3D_texture)=252
 sizeof(struct WAZP3D_blendstage)=18
 sizeof(struct state3D)=56
SetDrawRegion
setting the bitmap
 StateON(W3D_DOUBLEHEIGHT)=0
 yoffset=0
 context->width=640
 context->height=480
 WC->large=0
 WC->high=0
 SameSize=0
 WC->large=640
 WC->high=480
no scissor yet
 context->scissor.left=0
 context->scissor.top=0
 context->scissor.width=640
 context->scissor.height=480
 sizeof(union pixel3D)=106
 sizeof(union oper3D )=4
 sizeof(struct fragbuffer3D)=24
 sizeof(struct HARD3D_context)=68246
 sizeof(struct SOFT3D_context)=2525636
 sizeof(struct SOFT3D_mipmap)=3072
 sizeof(struct SOFT3D_texture)=36936
 sizeof(struct state3D)=56
SOFT3D_Start
 PrefsWazp3D=140614832
HARD3D_Start
SOFT3D_Start:OK
Bitmap size changed !!!!
SOFT3D_SetBitmap
 bm=139872440
 bmdata=270545792
 bmformat=5
Bitmap format changed
 bmformat=5
 bits=16
 Rbits=5
 Gbits=6
 Bbits=5
 Rpos=11
 Gpos=5
 Bpos=0
 ChangeOrder=0
SetImage to bitmap data
SetImage
 Image8=270545792
 x=0
 y=0
 large=640
 high=480
 bits=16
 offset=0
 SC->Image8=270545792
 SC->Image8=270545792
HARD3D_SetBitmap
large=640
high=480
startgl
pos: 0 0 size: 640X480
OS_StartGL
AROS: HC 149993382
AROS: HC->awin 140534648
AROS: mesawin <CoW3D V3> at 80 60 size 648X499
AROS: mesawin (borders 4 15 4 4 )
AROS: Cant create Mesa context
OS_CurrentContext
Image8=270545792
SetDefaults
SOFT3D_SetDrawState
 state:Changed1 UseTex0 ST0 ZMode7 TexEnvMode0 BlendMode33 UseGouraud0 UseFog0
 CurrentRGBA: 255 255 255 255
SOFT3D_Flush
HARD3D_Flush
OS_CurrentContext
---------
HARD3D_SetDrawFunctions
OS_CurrentContext
hc=149993382
state->UseTex=0
state->ST=0
state->gltex=0
state->TexEnvMode=0
state->ZMode=7
state->BlendMode=33
state->UseGouraud=0
state->UseFog=0
state->PerspMode=0
Gouraud OFF
depth OFF
Not UseTex
----------------------
SOFT3D_SetClipping
 xmin=0
 xmax=639
 ymin=0
 ymax=479
HARD3D_SetClipping
OS_CurrentContext
 context->width=640
 context->height=480
 context->depth=16
 context->bprow=80
 context->format=0
 context->drawregion->BytesPerRow=80
 context->drawregion->Rows=480
 context->drawregion->Flags=146
 context->drawregion->Depth=8
 context->drawregion->pad=24904
 direction=W3D_CCW
DoUpdate
SOFT3D_AllocZbuffer
 large=640
 high=480
HARD3D_AllocZbuffer
OS_CurrentContext
SOFT3D_ClearZBuffer
HARD3D_ClearZBuffer
OS_CurrentContext
CheckWarp3D:
==============================
 Wazp3D->drivertype=4
Hardware W3D_Driver available
==============================
========= W3D_Driver <Wazp3D - Alain Thellier - Paris France 2012 - Beta 52> soft:0 ChipID:1 formats:32767 =======
------------------------------
destformat:1

curious that it starts with error while continuuing on.. tells me nothing.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: Zac67 on September 14, 2013, 09:46:27 AM
.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 15, 2013, 03:19:27 PM
no comments? sigh, as usual the forum is good for phantasizing but talking some concrete solutions immediately drives everybody away. qed..
Title: Re: Bounty for better Radeon Mediator drivers
Post by: woof on September 16, 2013, 09:36:44 AM
Hello Wawa

>no comments?
Let me some time I was in week end :-)

>curious that it starts with error
This not an error value but the adresse of "error" that is a parameter for W3D_CreateContextTags
W3D_Context *W3D_CreateContextTags(ULONG *error, Tag tag1, ...)

The real error occurs here:
"AROS: Cant create Mesa context"
= Mesa/Gallium Error

>cow3d rendering to bitmap shows an empty window
That is very strange as the debug say

SOFT3D_Flush
 SC->FragSize2=28

That mean that (at least) 28 pixels were written to bitmap
so something should appear !

Perhaps there is a probleme with 16 bits modes. Does 24/32 bits works ?

I will try to download Aros68k and test cow3d myself

BTW can I send someone (Olaf?) the latests Wazp3D sources to update the Aros version to my latest build

Alain Thellier (Wazp3D)
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 16, 2013, 09:54:36 AM
yes, in case of cow3d something finally appears as i said, but only after you move the window around. i can guide you to setup working  aros 68k environment in case of problems i guess. what concerns the sources olaf is not an aros dev, nor me of course. do you know who commited wazp to aros contribs initially and who did the gallium backend? krzysztof(deadwood) perharps? ideally it sould be the same person ill ask around.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: Bamiga2002 on September 16, 2013, 11:47:47 AM
Yes, I'd be interested in Radeon W3D-drivers for OS 3.x.
(http://www.amiga.org/forums/showpost.php?p=721421&postcount=6)
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 16, 2013, 12:29:55 PM
unfortunately going aros way we will not receive 3.x compatible drivers, only aros68k drivers in best case. also i dont think radeons 9xxx would be supported. with all the software side ready at hand finding a gallium compatible 5v pci card would be another challenge. (preferrably a nvidia one afaik)
http://www.x.org/wiki/GalliumStatus/
http://nouveau.freedesktop.org/wiki/CodeNames/
Title: Re: Bounty for better Radeon Mediator drivers
Post by: OlafS3 on September 16, 2013, 04:10:05 PM
there are more and more developers interested to directly support AROS 68k so I will need AROS 68k as compiler target in future.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: yssing on September 16, 2013, 06:35:07 PM
@wawrzon Wouldnt aros 68k support be good enough?
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 16, 2013, 07:42:50 PM
Quote

@wawrzon Wouldnt aros 68k support be good enough?

that depends on everyones personal preferences. if aros 68k wass ready for prime time id say, yes. but since some serious issues remain i just want to warn about the actual situation, rather than spread promises.

since the thread is originally about radeon driver, another question would be what particula cards would come into question. i dont have definite answer to that unfortunately.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: vox on September 16, 2013, 08:45:46 PM
Quote from: kickstart;747698
Sometimes (many times) is like os4 is against the real amiga.

Well, I have been around for a lot of time, and I don`t have that impression. Its true Hyperion ceased any 68k development once moved to OS4 development, but they ceased to develop PPC games too - seems even OS4 is way too big task for such a small company.

Its easy to blame it when you look it like that, but to me AmigaOS 4 is as much Amiga as any AmigaOS 3.x is.

In terms of hardware, Classics have proven to be extremely limited at the end of 90s, so I always look forward for new generation of more open standard boards.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: vox on September 16, 2013, 08:49:52 PM
Quote from: OlafS3;747859
they want to sell their licenses and avoid everything that would support users on other platforms. I can remember reading a thread where someone asked one of the Friedens to compile something for another platform (I think it was Warp3D for 68k) and the answer was I could but I do not, buy AmigaOS when you want to use it (i do not remember the exact phrases). That shows the attitude...

As Wawa said only chance is AROS 68k. BTW the same is partly the case on MorphOS team too. Both "closed-source" and "commercial" camps want to defend their advantages and are not interested in sharing code in areas where they think they are better than the others. It is their right of course but it would have been better in a common interest to work together. But that is not the case (and will propably never be).

Of course it is different outside the "core" and even in the teams there are exceptions that are very open (like Fab from MorphOS-Team) but I think every technology that is seen as advantage like MUI4, P96, Cybergraphics 5 and so that is copyrighted will not opened or at least cross-compiled.

Now I understand it better. Its because new developed software points that have pushed MOS and OS4 forward will never get 68k compiled, and you are advised to move PPC, which you dislike.

Well, since those components are only things that do make a difference and move ahead, its understandable why they will not do it.

To be honest, would also like its one great (open source or not) OS, be it 68k,PPC or x86 (or all tree) but that is unlikely to happen.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: OlafS3 on September 16, 2013, 08:53:29 PM
Quote from: wawrzon;748133
that depends on everyones personal preferences. if aros 68k wass ready for prime time id say, yes. but since some serious issues remain i just want to warn about the actual situation, rather than spread promises.

since the thread is originally about radeon driver, another question would be what particula cards would come into question. i dont have definite answer to that unfortunately.


"i just want to warn about the actual situation, rather than spread promises."

Yes? Example?
Title: Re: Bounty for better Radeon Mediator drivers
Post by: nicholas on September 16, 2013, 10:19:16 PM
Quote from: vox;748139
Now I understand it better. Its because new developed software points that have pushed MOS and OS4 forward will never get 68k compiled, and you are advised to move PPC, which you dislike.

Well, since those components are only things that do make a difference and move ahead, its understandable why they will not do it.

To be honest, would also like its one great (open source or not) OS, be it 68k,PPC or x86 (or all tree) but that is unlikely to happen.


Your wish already exists in the form of AROS and at some point in the future both MorphOS and OS4 will cease to be developed. When that will happen is unknown but it will happen and at that point the code for AROS will still be there for anyone who cares to do whatever they want with on whatever hardware they want.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: XDelusion on September 16, 2013, 10:47:51 PM
Quote from: nicholas;748145
Your wish already exists in the form of AROS and at some point in the future both MorphOS and OS4 will cease to be developed. When that will happen is unknown but it will happen and at that point the code for AROS will still be there for anyone who cares to do whatever they want with on whatever hardware they want.


Maybe so... but it's going to be a LONG wait the way things are going.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: nicholas on September 16, 2013, 11:02:53 PM
Quote from: XDelusion;748147
Maybe so... but it's going to be a LONG wait the way things are going.


True. Very true. :/
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 17, 2013, 12:37:11 AM
Quote from: vox;748138

In terms of hardware, Classics have proven to be extremely limited at the end of 90s, so I always look forward for new generation of more open standard boards.


this is not the subject, whether or whatever the limits are these are concerns os amiga/aros68k users, standard boards are available for other aros users rather than what you choose to use. no need to lecture us who know better.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 17, 2013, 12:42:36 AM
Quote from: OlafS3;748141
"i just want to warn about the actual situation, rather than spread promises."

Yes? Example?


what are you upset about? as winuae user its not your concern, while those who might expect to use some sort of pci rtg boards in their amiga setups will have to make up their mind about the issue (im not an expert unfortunately). therefore i just propose a solution that the more knowledgeable might or might not find feasible. i am not sure if and which a nvidia card will work in (say) a mediator. none has ever tried so far i know.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 17, 2013, 12:44:08 AM
Quote from: vox;748139
and you are advised to move PPC, which you dislike.

im bored with your agenda, to tell the truth..
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 17, 2013, 12:53:46 AM
Quote from: woof;748093

BTW can I send someone (Olaf?) the latests Wazp3D sources to update the Aros version to my latest build

Alain Thellier (Wazp3D)


i have now the answer from the developer, who commited wazp to aros. i will establish the contact or pass the code over when necessary, but perhaps we dont need a forum for that. alain do you still have my mail? otherwise ill contact you. also whatever support you need and can expect from me im ready to provide;)
Title: Re: Bounty for better Radeon Mediator drivers
Post by: woof on September 17, 2013, 09:27:18 AM
Yes it was Matthias Rustler that dit the Wazp3D Aros port
I will contact him soon (and you Wawa) but I need to check the latests sources and I dont have the time today
But I will do that this week
Same  for aros 68k iso + cow3d : dont have the time yet but I will check that soon

Alain Thellier (Wazp3D)
Title: Re: Bounty for better Radeon Mediator drivers
Post by: wawrzon on September 17, 2013, 12:28:23 PM
alright, alain!
Title: Re: Bounty for better Radeon Mediator drivers
Post by: OlafS3 on September 17, 2013, 01:40:01 PM
Quote from: wawrzon;748165
what are you upset about? as winuae user its not your concern, while those who might expect to use some sort of pci rtg boards in their amiga setups will have to make up their mind about the issue (im not an expert unfortunately). therefore i just propose a solution that the more knowledgeable might or might not find feasible. i am not sure if and which a nvidia card will work in (say) a mediator. none has ever tried so far i know.

I was only wondering... one said AROS 68k is "unusable" and you talked a little negativ too so I was wondering what is missing. Functionality (when using Magellan) certainly is competitive to any other 68k platform. Of course supporting PCI/Zorro cards on real hardware is out of my scope.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: vox on September 17, 2013, 06:02:27 PM
Quote from: wawrzon;748163
this is not the subject, whether or whatever the limits are these are concerns os amiga/aros68k users, standard boards are available for other aros users rather than what you choose to use. no need to lecture us who know better.

It wasn`t intended to be a lecture, but personal sentiment and attitude.

PowerPC route was chosen because 68k seemed dead. But this were times when everything was hardware tied. Today, with AROS 68k, FPGA machines and UAE almost universally avail, 68k exists on both software and hardware level, even in better position then ever. 68k code we can run on AmigaOS 4 and MorphOS too ... and its most present "Amiga"

Back on topic, if it would help guys using the Classics, would donate to such bounty.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: slaapliedje on November 19, 2013, 05:30:21 AM
Being an owner of both a Radeon board and a Mediator, I would totally help with this bounty, though in reading the whole thread, I realize that's not going to ultimately help.

I also wish they (Elbox) hadn't strong armed Ratte.  I was following his development for a bit, then kind of faded away for a bit (got busy) then when I came back to check on it, the driver had disappeared and the whole crap had hit the fan.

I really do think the future is AROS (68k or otherwise) though, being a huge Open Source advocate helps.  

On the note about wawrzon mentioning support for the nVidia cards.  I believe at the very least that the earlier Radeon cards (7200 - 92xx) were one of the first fully supported chipsets from the earlier hardware accelerated mesa efforts.  I recall it being nvidia-glx, then utah-glx (probably have those backwards), then radeon had full support.  After that the OSS implementations seemed to dry up as both ATI and nVidia ended up getting more proprietary (licensed?) technologies in their cards, and so more closed specifications.  Matrox (MGA) cards could be lumped in with that group as well.

The real question is, do Matrox and / or Nvidia make 5v PCI cards that are easily available to be used in a Mediator (assuming drivers were made)?

As a last note, I also must state that Elbox... WTF?  We have to buy your hardware, does it really matter who makes the drivers for it?  I even ended up buying a Spider 2.0 and a Radeon (which I ended up not using, and use one that has more memory on it).  It's a shame that crap like this still goes on, but hey it is the Amiga community.  Maybe I haven't been paying attention, but I don't see this much back stabbing and legal crap creep up in the Atari community.

slaapliedje

P.S. I realized I necro'd this thread a bit, but it always kind of fits in an Amiga forum, I figure a few months isn't terrible.  :D
Title: Re: Bounty for better Radeon Mediator drivers
Post by: klx300r on April 21, 2018, 10:26:53 PM
a 5 year bump from this thread ;) http://www.amiga.org/forums/showthread.php?t=73448
Title: Re: Bounty for better Radeon Mediator drivers
Post by: Iggy on April 22, 2018, 04:18:36 PM
Argh! I'm not going to LIVE long enough for you die-hard fanatics to get this "just right".

So, yes, I moved to NG, and as that transitions I'll follow that.

Because my ECS and AGA hardware is hopelessly dated, and your FPGA based solutions will never be as powerful as dedicated silicon.

So I should plug my legacy hardware into a Mediator, limiting my expansion options to PCI, then plug an FPGA in as a processor replacement (and at least partially, a chipset replacement)?

And I get a slower system that can't use recent gpus? After I've essentially jacked the whole system up and driven something underneath it to replace it?

And you guys make fun of the cost of a Aeon board?
Keeping in mind that I use a Mac (although I'm open to that option)?
After you've spent HOW much?

Just so you have "the real thing", which it isn't since the things that really accelerate and expand the system are third party?

Truly cuckoo.

Understand, I'm in favor of better Radeon drivers, with a qualifier.

The term Radeon for legacy ends at the 9200.

And the OS4 drivers at least provide acceleration for that, OS3 doesn't have that.

Also, I haven't been worried about W3D acceleration since I moved to a 9800 (which OS4 has a rather limited driver for), but W3D is still a feature of latter OS4 video drivers (for those of you who are truly committed or "commit-able" anyway).

Just my opinion, but the foot dragging of the legacy fanatics is the primary reason it hasn't gone anywhere. Not available resources, the users.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: billt on April 23, 2018, 02:39:32 AM
Quote from: Iggy;838684
The term Radeon for legacy ends at the 9200.



Ask Elbox for a Mediator with PCI-express slots. We'll never get to their bandwidth capability, but at least o physically plug a newer card in...
Title: Re: Bounty for better Radeon Mediator drivers
Post by: Iggy on April 23, 2018, 03:45:06 AM
Quote from: billt;838698
Ask Elbox for a Mediator with PCI-express slots. We'll never get to their bandwidth capability, but at least o physically plug a newer card in...

I can't picture Elbox having the engineering competence need to implement PCIe, although some kind of bridge could be used to at least provide a basic PCIE compatible slot.

Bandwidth? No, legacy Amiga's wouldn't exploit it, but at least something more modern than the R200 based cards could be supported.
Then again, as a graphics upgrade, the R200 is still quite a bit more powerful than the Amiga's own chipsets built in graphics.

So...sorry to divert the thread.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: Karlos on April 23, 2018, 07:16:51 AM
In had to laugh at the bandwidth comment. Maybe a GRexx in an 060 powered A4000 might get close to that the R200 is accustom to. Nevertheless I believe the point is more about recency. Its hard to find PCI Radeon cards these days. Most of the R100 and R200 cards were AGP.
Title: Re: Bounty for better Radeon Mediator drivers
Post by: Iggy on April 23, 2018, 10:12:35 AM
Quote from: Karlos;838703
In had to laugh at the bandwidth comment. Maybe a GRexx in an 060 powered A4000 might get close to that the R200 is accustom to. Nevertheless I believe the point is more about recency. Its hard to find PCI Radeon cards these days. Most of the R100 and R200 cards were AGP.


Yep, I think the most recent PCI card I have come by was HD5450 based. And they are hard to find.
And PCI 9200s still pop up on eBay.