OpenGL 3.1 Not Likely In Mesa Until 2013

[Kivada]

I can't understand it. Mesa has so great support from Intel, Red Hat and others big giants of IT, and they couldn't implement 3 years old specs, where nVidia and AMD could do it in so short time. What is the main reason? No so much ppl or money for work?

Allot of the problem is the fact that the hardware itself lagged years behind the spec, take a look, Intel GPUs didn't even support OpenGL3 till the Sandy Bridge era: https://en.wikipedia.org/wiki/Compar...ocessing_units

The new HD Graphics 2500 and HD Graphics 4000 GPUs can now handle OpenGL4, but there is no reason to implement it on them as they are far too slow to handle what most games are doing these days

Intel's team seems to be one of the major driving forces behind MESA's OpenGL code as AMD's team is too small devote the manpower to add to the spec seeing as they have such a large back catalog of GPU hardware to get working on the existing MESA code before they can even think about adding advanced features to the OpenGL stack. You've gotta get the hardware running in the first place before you can make it stable, fast and add features else it becomes a half assed monstrosity.

[orome]

mesa desperately needs someone with deep pockets to back it up

I hope not, beacues it would be a failure for OSS community. The idea was that payed developers provide kickstart and guidance for idependent contributors, but it looks like there is a problem with this. Can it be that mesa has so high requirements and steep learning curve (knowledge of 3d stack and hw), that the pool of possible contributors is so limited?

<troll>
With so many people on this forum that know what to do and how to do it better, I would expect no shortage of developers and high quality testers.
</troll>

If this project really needs more corporations to back it up it would be more of an industrial collaboration than community oss, and that is not a novel concept.

[Delgarde]

I can't understand it. Mesa has so great support from Intel, Red Hat and others big giants of IT, and they couldn't implement 3 years old specs, where nVidia and AMD could do it in so short time. What is the main reason? No so much ppl or money for work?

Hint - Nvidia and AMD (or at least, the part that used to be ATI) are specialists in creating graphics cards and drivers. It's how they make money.

[Kivada]

Hint - Nvidia and AMD (or at least, the part that used to be ATI) are specialists in creating graphics cards and drivers. It's how they make money.

Fun Fact: Nvidia doesn't give a shit about open source. Nouveau is an independent team and have no backing from Nvidia.

Fun Fact: AMD's open source team is still building their driver infrastructure for a decade's worth of GPU FAMILIES, maybe they could devote the man hours needed to implement more recent revisions of the OpenGL stack if they hired another 10-20 devs.

[Delgarde]

Fun Fact: Nvidia doesn't give a shit about open source. Nouveau is an independent team and have no backing from Nvidia.

Fun Fact: AMD's open source team is still building their driver infrastructure for a decade's worth of GPU FAMILIES, maybe they could devote the man hours needed to implement more recent revisions of the OpenGL stack if they hired another 10-20 devs.

I know. Nevertheless, the question was why Nvidia and AMD can do a better job of writing drivers than the larger companies who contribute to Mesa. And that's the answer - they're the ones who build the hardware, the ones who have the documentation, the ones who know what they're doing. It would be astonishing if they couldn't do a better job.

[curaga]

Well, didn't Intel's Tampere office alone have over 30 graphics people? Though I heard they're mainly working on PowerVR stuff...

[89c51]

I hope not, beacues it would be a failure for OSS community. The idea was that payed developers provide kickstart and guidance for idependent contributors, but it looks like there is a problem with this. Can it be that mesa has so high requirements and steep learning curve (knowledge of 3d stack and hw), that the pool of possible contributors is so limited?

<troll>
With so many people on this forum that know what to do and how to do it better, I would expect no shortage of developers and high quality testers.
</troll>

If this project really needs more corporations to back it up it would be more of an industrial collaboration than community oss, and that is not a novel concept.

most, if not all, successful FOSS projects have companies that back them or at least have a way of making money. For mesa the support it gets from Intel AMD Red Hat etc is not enough sadly. its a huuuuuge project that requires far more people in order to be able to compete.

i don't know if they can do more in terms of resource management but even devs have commented several times that they need people. and you won't get them easily without money.

[89c51]

Fun Fact: Nvidia doesn't give a shit about open source. Nouveau is an independent team and have no backing from Nvidia.

Fun Fact: AMD's open source team is still building their driver infrastructure for a decade's worth of GPU FAMILIES, maybe they could devote the man hours needed to implement more recent revisions of the OpenGL stack if they hired another 10-20 devs.

i don't think that AMD ever said that they will write the driver for the community. they just said they will provide documentation. the fact that they actualy do employ some people is really nice of them.

and yes many companies could pay someone to write them an OpenGL state tracker or whatever is needed. they don't see any benefit from it sadly.

[V!NCENT]

Don't rush it. Every OpenGL bump is improvement. Second hand/older OpenGL books are cheaper. Not a matter of just mastering OpenGL x and OpenGL y is already at the front door...

I like the fact that my Linux desktop is stable and clean; keep it that way!

[brent]

I don't think this is a big deal. OpenGL 3.0 support barely works, and is still full of bugs. It's definitely a good idea to wait for OpenGL 3.0 support to settle and stabilize before leaping ahead to the next version as quickly as possible. The next real milestone is OpenGL 3.2 anyway. Geometry shaders!