Open ATI Specifications For R100-200

[phoronix]

Phoronix: Open ATI Specifications For R100-200

On Friday we talked about ATI preparing to release more GPU documentation to the public and without any Non-Disclosure Agreement, and today we have a few new details regarding the specifications for earlier Radeon GPUs. AMD's John Bridgman has posted in this Phoronix Forums thread that he is trying to re-release the GPU specifications for the R100~200 generation of graphics processors.

http://www.phoronix.com/vr.php?view=NjIwMg

[oliver]

It's posts like these that give me shivers down my spine, good ones mind you.

How small of a step it is, it is a step, I think it's a leap, into the right direction. Too little to late? I don't think so. Those specs where (partially?) released years ago, and even now, today, they are putting effort into bringing these out.

Makes me happy to be sticking with ATi/AMD for a few more years, even though all the hype is about those other fella's at the moment.

[Uchikoma]

When it actually happens, that's when I'll start bouncing off walls.

But for now, I'll be reserved in my praise and joy, and say that I'm really, really glad that they (appear) to be serious about helping open source development of their graphic drivers.

[pawstar]

Do you think we will get the specs for the All-in-Wonder series? So far, my luck has been abysmal in trying to get them to work - and I know that I am not alone in this. It would be great to use 'em for compact home theater setups.

[Thetargos]

Problem with All-In-Wonder products AFAIK is the use of third party chips for which ATI/AMD has no right to give away the specs of. Please someone prove me wrong, as indeed AIW products would really make some nice HTPC GPUs

[yoshi314]

i've seen the original post before. good stuff.

what kept them so long? this should have happened as soon as r500 chips hit the market. maybe even as soon as r300 were released.

i wonder what will be there that's still missing from older specs that dri devs have got.

[chrisr]

Full support for HyperZ, please.

I always seem to be stumbling across issues with the R100 and R200 drivers where I am later told "You need to disable HyperZ" before it will work properly. Which is apparently the reason why HyperZ isn't enabled by default in the first place. However, HyperZ gives a noticeable increase in FPS for (say) glxgears, so I would really like to be able to use it properly.

[Extreme Coder]

But,
what use will these documents have for the drivers? I very much doubt the oss X.Org devs will rewrite the R100-R200 drivers. But could this lead for performance like the fglrx driver?
Because for example NWN was unplayable with the OSS driver, while it was playable with fglrx (and at good framerates too). And lots of other games too. (on ATI Radeon 9200 SE)

[bridgman]

>>I very much doubt the oss X.Org devs will rewrite the R100-R200 drivers.

We don't expect that any of the existing drivers will be rewritten as a direct result of this either. The goal is to reduce the amount of trial-and-error work (which sucks up a lot of developer time), to help with stability issues caused by missing information, and to enable some additional functionality.

On the other hand, as events come along which *do* justify a driver rewrite (Gallium, DRI2, TTM, kernel modesetting etc...) we do expect that this initiative will make the rewrite efforts easier and more successful.

[Berniyh]

>>I very much doubt the oss X.Org devs will rewrite the R100-R200 drivers.

We don't expect that any of the existing drivers will be rewritten as a direct result of this either. The goal is to reduce the amount of trial-and-error work (which sucks up a lot of developer time), to help with stability issues caused by missing information, and to enable some additional functionality.

On the other hand, as events come along which *do* justify a driver rewrite (Gallium, DRI2, TTM, kernel modesetting etc...) we do expect that this initiative will make the rewrite efforts easier and more successful.

I think, that we have already seen what it would do.
Right after the release of the first set of specs there were a lot of changes, which corrected a lot of things.
Of course that driver was in a different shape than the quite stable ati/radeon drivers are, but I guess that there would be a similar effect.

In addition, as far as I know, some thinks don't work with the open source drivers yet. For example the Powermode setting (aka Powerplay).
For this kind of stuff I hope the specs would be useful.