Amiga.org
Amiga News and Community Announcements => Amiga News and Community Announcements => Amiga Software News => Topic started by: Ratte on December 02, 2012, 06:55:27 PM
-
http://www.a1k.org/forum/showthread.php?t=35144
mediator (bigbox/z3)
prometheus
amithlon
-
Grrrrr... must set up Amithlon machine to test this with!!!
-
Can I have 250MB gfx mem +6MB shared PCI mem?
-
Great news :-)
-
Can I have 250MB gfx mem +6MB shared PCI mem?
According to another thread I've been in, the 128MB issue is due to the way the Radeon 92xx's memory banks are set up. Instead of one bank of 256mb, it's in two different 128mb banks.
With the Mediator drivers, you'd need to specify the Radeon Memory in the ENV-Arc:Mediator:RadeonMem file to be 15 (for full on resolutions, I don't think any resolution really uses more than that) and then use AddMem to add the second bank of Memory. I think I had it working at one point, but then reinstalled, and didn't get it working since... gonna have to reinstall again, since I think something got corrupted along the way :( I keep having stupid lockups in all Radeon resolutions.
Just haven't had the time (or focus) to be able to do it. Probably will this coming weekend.
But I'm wondering if using this OpenPCI driver will work better...
-
AddMeming the 2nd bank of 128MB would cause it to be normal Amiga "fastmem" I would think.
So this means Elbox drivers only support 128MB gfx mem with their drivers. :(
And probably this new OpenPCI driver has same limitation. :(
> u don't need more than 128MB
Yes I do need 200MB or more of gfx card mem.
If there is only 128MB then first of all I have to try to cram all my gfx in there which don't fit.
And I must include lame instructions to turn down your workbench before starting the game. And to quit all other programs that are running on your gfx card because they are eating to much memory. Ugh.
I wish Elbox would have just explained this on their website.
-
This driver can be used on ati 7000 - 7500 pci and agp cards ?.
The 7000 needs 3.3 volts or i can just install it on the prometheus ?
-
So this means Elbox drivers only support 128MB gfx mem with their drivers. :(
Yes I do need 200MB or more of gfx card mem.
I wish Elbox would have just explained this on their website.
The drivers(s) support more than 128MB.
BUT ... there are different Radeon 256MB designs out there.
Some of them have 256MB in ONE block, others came with 2 parts a 128MB.
P96 is only able to handle linear memory (one block).
So its not a Elbox-radeon or xyz-radeon-driver problem ... P96!
V0.99 limitation to 128MB is only until i changed some code to place the spritegfx at the beginning of memory .. the sprite-pointerregister is only able to adr. 128MB also on oneblock 256MB radeons.
@soviet: R100 is not supported, only R2xx (8500AGP - 9250PCI).
-
The drivers(s) support more than 128MB.
BUT ... there are different Radeon 256MB designs out there.
Some of them have 256MB in ONE block, others came with 2 parts a 128MB.
P96 is only able to handle linear memory (one block).
So its not a Elbox-radeon or xyz-radeon-driver problem ... P96!
V0.99 limitation to 128MB is only until i changed some code to place the spritegfx at the beginning of memory .. the sprite-pointerregister is only able to adr. 128MB also on oneblock 256MB radeons.
@soviet: R100 is not supported, only R2xx (8500AGP - 9250PCI).
Thanks ratte so i have to get a 9200 now, and do some kind of volt mod the prometheus doesn't have 3.3 volts
-
So its not a Elbox-radeon or xyz-radeon-driver problem ... P96!
I bet CGX has the exact same limitation. :)
-
The Prometheus open sourced and this too! Very good timing.
I need to do a voltage mod on the Prometheus and replace that hot voodoo card.
Good work!
-
Works here with my old Prometheus :-)
If you use an ATX PSU it's quite easy to get 3.3v line and connect it to a pci-riser so you don't have to risk your dear prometheus/graphic cards too :-)
-
Hello,
It is just the copy of my post on amigaworld.net:
Two weeks ago I read on a1k.org that two people - MastaTabs and Ratte -- made available OpenPCI driver for the Radeon gfx cards.
I was a bit surprised that they made a driver for exactly the same ATI chipsets, which are just supported by Elbox Radeon driver. There are so many other gfx cards around...
I looked into the openpci driver code to check the name of its author but found that there isn't any copyright info inside. Nobody signed this driver. It seemed to me a little suspicious because developers who spend many weeks or months on such a project usually sign their code.
I sent the archive with this driver to Pawel Stypula - author of the Elbox driver for the Radeon gfx cards. He is a very busy man but promised to make a look into it as soon as has a moment of a spare time.
I just received his response. He is totally shocked. More than 90% of our Radeon related code was directly copied from our Radeon.card driver (ver. 2.12 from 09 Sep 2010) to the openpci Radeon.card driver. Almost all functions were taken from our driver and inserted by simple "copy and past" to this "new" driver. Many of the procedures were taken without changing even single line of the code. The only actual change that has been made in the code of functions taken from our driver was removing these parts of code, which are related to Mediators for A1200.
The blocks of the code taken from our driver without altering even one byte in them are so big as several hundreds of processor opcodes! There are over a dozen of such big directly copied blocks of code. The biggest one has 637 assembler lines (1898 bytes of the driver code)!
I'm really surprised that someone decided to make an almost 1:1 copy of our code, inserted it as an essential part of his driver and announced it as his work.
The development of our Radeon driver took months of work in 2007 and 2008. This was a first support for any ATI Radeon gfx cards in Classic Amigas. There were many undocumented Radeon issues which had to be solved in the code of this driver. Solving them required many time and hard work.
Nobody wants their work being stolen, we too. If somebody wants to use our code he should contact us for the license. Using our copyrighted code without our permission is illegal.
The driver released by MastaTabs and Ratte is clearly illegal as containing the stolen code.
Best regards,
Darek Smietana
ELBOX COMPUTER
Postscriptum:
It is just the removal of checks on which Mediator (A1200 or ZIII Amiga) and in what mode (MMU / NOMMU) driver is working what results in the slight increase in speed tests measured by P96speed.
-
It's a shame that they weren't honest about the source of the code, but at least it's free.
-
statement:
http://www.a1k.org/forum/showthread.php?p=585102
-
statement:
http://www.a1k.org/forum/showthread.php?p=585102
Dear MastaTabs & Ratte,
The fact that many Mediator drivers are available for downloads at our website has nothing to do with their commercial status. In the archive enclosing our Radeon.card driver there is a disclaimer which clearly informs that our driver is a commercial software and is copyrighted.
As in the code of the OpenPCI Radeon.card driver there are many, many big fragments of code copied/stolen 1:1 from our driver the OpenPCI driver is illegal and as such should be immediately removed from every places.
It is completely unimportant if you sell this illegal software or you give it away for free. Our Radeon.card driver was written for Mediators and no one contacted us for licensing its code to any other hardware solution.
Imagine what would happen to you if you do the same with any software which works with Apple computers only. If you copy big fragments of Apple copyrighted code, remove protection (which join the code with Apple computers hardware) and you distribute such code for free for users of any PC computers...
Your suggestions that we based the Radeon initialization code on the Linux source is completely wrong. This code was written by Pawel Stypula long time before ATI released the AtomBIOS parser source code. The first working version of the code for initialization the Radeon gfx card we have had in November 2004. This code was firstly working on the Radeon 9250 AGP card installed in the Dragon 1200 busboard. We still have the sources of our earliest versions of the Radeon.card.The Radeon initialization procedures we based on the reverse engineering of x86 BIOSes of the Radeon cards - it was really hard work.
From the beginning the code of our Radeon.card driver has been written in the 68k/Coldfire assembler. Not a single line of this code is written in C+.
As in the OpenPCI Radeon.card driver the main part of initialization procedure (the whole AtomBIOS parser) is the exact copy byte-by-byte (over 600 assembler lines of code) taken from our driver do not lie that you based it on any other source than our driver. This code is simply stolen as many, many other parts from our driver code.
Our Radeon.card driver surely includes the parts of code of Voodoo.card driver as a framework. The parts of our Voodoo.card driver, which was the first (released in November 2000) driver for Voodoo3 gfx cards working with the P96 gfx system.
Please remove immediately your illegal driver from any downloads.
Best regards,
Darek Smietana
ELBOX COMPUTER
-
GRex (4000) + Radeon
http://www.a1k.org/forum/showthread.php?goto=newpost&t=45449
-
Listen guys I'm not a developer and I'm not going to get into this but can you please stop bickering on a public forum about allegations of someone copying your code?
There are courts for a reason, if you think somehow this is worth pursuing, involve the courts or carry it over direct PM. At this point everyone bickering about it is just making themselves look like a bunch of kids.
-
GRex (4000) + Radeon
http://www.a1k.org/forum/showthread.php?goto=newpost&t=45449
What is performance of the driver like under Amithlon compared to native Linux drivers?
-
@TeamBlackFox
Look at the posting dates. :)