Warp3d Voodoo vs. Radeon ClassicOS

[Motormouth]

@amiadudeorwat and trixster

It is obvious that I need to keep up with the EAB amiga sonnet thread........

@Iggy

I looks like the other side of the pond wants an R200 driver too.  

[kev711]

not my work that is hedeons work i just posted cause not 1 has noticed it on youtube as its been there for awhile lol
just get in touch with him as it still needs work
im just the hardware guy

[Karlos]

Porting the OS4 drivers for R100/R200 to classic is a bit more involved than potentially writing from scratch, in part because they share an OS4 only resource with the compositing engine. Occasionally there are other obstacles too, for the unwary. For example, the Picasso 96 driver for Permedia had to be changed to get a functuonal Warp3D driver for OS4 since the initial 2D subsystem only cared about minimal alignment of bitmaps and not the specific encoded sizes required for hardware rasterizarion to work. This meant most bitmaps allocated by the graphics system were not suitable for 3D.

[Karlos]

In hindsight, I'm also of the opinion that a less modular warp3d implementation for classic makes sense. For example, having a whole array of precompiled warp3d libraries and you install the one specific to your rtg/3d chip.

This may sound like heresy and a huge retrograde step but it has a lot of advantages for older systems. For example, in warp3d the rtg library component handles allocation of memory and the 3d component has no say, but on an 8MB card the allocation strategy that works best will nor be the same as some 128MB one. Also, it would remove many layers of indirection and partial features like multiple card support that add a lot of complexity. Finally, just support v4 and v5 vertex arrays and emulate the v3 calls using them. A single drop in library that will either work or fail to open if it isn't correct for your system.

[Motormouth]

In hindsight, I'm also of the opinion that a less modular warp3d implementation for classic makes sense. For example, having a whole array of precompiled warp3d libraries and you install the one specific to your rtg/3d chip.

Reminds me of the SGI Indy that I use to have at work:)

[Motormouth]

I just read this ~ four year old thread:

http://www.amiga.org/forums/showthread.php?t=65900

Sniff, Sniff , it make me want to cry

[klx300r]

@ Motormouth

bounty time again then:hammer:

[Karlos]

As I see it the problems are:
* Finding the people that have the time, skills and inclination. Writing hardware drivers is time consuming and a pain in the rear to debug.
* Getting access to the legacy W3D source. It has moved on since OS4.
* Identifying all the other problems. As I said before, even the 2D drivers for OS4 use the 3D engine for compositing so you know the hardware is initialised correctly. Its likely that a legacy R200 driver will need more set up as it can't depend on anything being done for it.
* Your new 68K R200 driver may suffer memory management problems without work to the rtg driver component. that in turn many need changes to Picasso driver.
* You may have even more problems if your card is in a Mediator1200, at least from the software side.
* Maybe GRexx owners want someblove too.
* And finally, doing all this without elbox wanting to kick you where it hurts, because apparently they don't like other people writing drivers for their stuff.

If you think a bounty can address these then great.

[Iggy]

Drivers for later cards under OS4 have to use 3D gpu functions since there aren't any 2D functions.

[Karlos]

Drivers for later cards under OS4 have to use 3D gpu functions since there aren't any 2D functions.

My point is that the OS4 Warp3D driver for R100/200 relies on functionality nor available to OS3.x, so a back port becomes a bit more involved.

[nicholas]

Drivers for later cards under OS4 have to use 3D gpu functions since there aren't any 2D functions.

Ahem.... I'll just leave this here......

https://www.google.co.uk/search?q=karl+site%3Awww.amigaos.net

[Iggy]

My point is that the OS4 Warp3D driver for R100/200 relies on functionality nor available to OS3.x, so a back port becomes a bit more involved.

So it would be best to do it from scratch. Got it.

[Iggy]

Ahem.... I'll just leave this here......

https://www.google.co.uk/search?q=karl+site%3Awww.amigaos.net

Wow, it takes a lot of coders to be second best. :hammer:

[Karlos]

Wow, it takes a lot of coders to be second best. :hammer:

Taking part is what matters.

[Iggy]

Taking part is what matters.

Yeah, I know.
Hope the AmigaOS diehards don't get offended by the jab.

And outside of OS development, the base of developers creating useful programs isn't bad.

[Karlos]

I think I stand by my earlier point. A single warp3d library replacement, pre compiled for a specific GPU/PCI/rtg combo is the only way to get past all the issues, technical and political. The source can and should be as modular as you like, but you need a fresh start implementation wise. A driver for warp3d on the classic just isn't likely to happen. As it stands, reimplementing the API in a totally separate project doesn't violate anything and already happened with wazp3d. Use OpenPCI to get around elbox's self destructive desire to crush any third party drivers and open source the whole thing.