Guide to Get the Best Performace from the GMA 500

[lucazade]

yes, thank you, Lucyd was a disaster so yes, I went for karmic given that it is supposed to have the best performance with GMA 500. but I havent been able to make the graphics work. I also happen to have the latest kernel 2.6.31-22-generic and I read some solutions wont work with that. Do you or anybody have any idea what can I do?

https://wiki.ubuntu.com/HardwareSupp...oCardsPoulsbo/
Here you will find instructions for Jaunty/Karmic/Lucid.


Keep in mind Gma500 drivers are the same in all these releases and if you have a problem please explain it better otherwise it is difficult to help you.


- The post you quoted from sammyboy405 is for Jaunty, not Karmic.

- Karmic and Lucid performs in the same way, only difference is that in Lucid Xv video playback is broken (but there are workarounds reported in previous posts)

- The latest kernel of Lucid (2.6.32-24-generic) works well with Gma500, no issues.

[jbernardo]

The more I look at the binary libs included, the bigger mess this EMGD driver is turning up to be. Besides the insulting remark on Intel's FAQ that "Intel graphics chipset customers would not need access to the source code" which really should not applying to us end users stuck with poulsbo netbooks, they are packaging GPL licensed libraries without the respective source - libva.
Is Intel trying to **** off every linux user, or only us?

[simplygades]

Taking for granted that they do not offer support for their hardware by themselves, I can't think of a sane reason why a hardware company would not provide some kind of help to end-users trying to make it's hardware work. Or is it that Linux-Poulsbo combination (at least) disappointed users is an insignificant percentage to them?

[jbernardo]

I think the marketing/lawyer/PHB in charge at Intel thought that we would go away if they kept ignoring us, and that we wouldn't make enough of a fuss. Just like Oracle will soon enough find that their actions against Java will make a lot of sysadmins consider alternatives, also Intel will some day find out what are the consequences of their actions. My next netbook will probably be a AMD+ATI one, what about yours?

[simplygades]

I think the marketing/lawyer/PHB in charge at Intel thought that we would go away if they kept ignoring us, and that we wouldn't make enough of a fuss. Just like Oracle will soon enough find that their actions against Java will make a lot of sysadmins consider alternatives, also Intel will some day find out what are the consequences of their actions. My next netbook will probably be a AMD+ATI one, what about yours?

Personally, I swore not to buy anything with their logo on it from now on. Performace will always be sufficient for what I do, so what remains is ethics. Whatever the brand, I always prefer companies that don't take for granted that customers are a bunch of unthinking, ever-consuming creatures. I have a list of proposals for what to do with their remaining chips, when this time hopefully comes, but I can't write it here.

[yvesdm3000]

The more I look at the binary libs included, the bigger mess this EMGD driver is turning up to be. Besides the insulting remark on Intel's FAQ that "Intel graphics chipset customers would not need access to the source code" which really should not applying to us end users stuck with poulsbo netbooks, they are packaging GPL licensed libraries without the respective source - libva.
Is Intel trying to **** off every linux user, or only us?

When you look at the meego driver and at the driver from Imagination (makers of PowerVR), you'll notice how much they are the same. The Imagination driver is not an xorg driver (only OpenGL and direct framebuffer access), Intel simply used the libraries from Imagination and built a driver around it. I guess Intels hands are bound from any angle...

-Yves

[lucazade]

I guess Intels hands are bound from any angle...

Intel could buy Imagination in a fraction of second and without noticing. Am I wrong?

[jbernardo]

When you look at the meego driver and at the driver from Imagination (makers of PowerVR), you'll notice how much they are the same. The Imagination driver is not an xorg driver (only OpenGL and direct framebuffer access), Intel simply used the libraries from Imagination and built a driver around it. I guess Intels hands are bound from any angle...

-Yves

Intel at least could have provided us with the source for what they did on top of the Imagination driver, as well as pointing out which of the libs they used are GPL, are where we can get their source. This second part isn't even being nice - it is mandatory by the GPL.
And Imagination always claimed that the problem with open-sourcing the gma500 drivers is with Intel - if Intel provides their changes, and points back to Imagination to provide the originals, that claim would be emptied quite fast. As is, we got something even worse than the PSB drivers, a bunch of binaries without even the source for the 2D driver, and Intel removed the source they had already made available on the Meego repositories. I think I might still have it around, so if I have time I'll check if it builds and if we can separate the open from the closed source in the EMGD driver.

[yvesdm3000]

Intel could buy Imagination in a fraction of second and without noticing. Am I wrong?

I think there would be other parties not happy with that. ARM for sure, Qualcomm, Texas Instruments and others. It's mostly because Imagination is not for sale I guess...

-Yves

[yvesdm3000]

Intel at least could have provided us with the source for what they did on top of the Imagination driver, as well as pointing out which of the libs they used are GPL, are where we can get their source. This second part isn't even being nice - it is mandatory by the GPL.
And Imagination always claimed that the problem with open-sourcing the gma500 drivers is with Intel - if Intel provides their changes, and points back to Imagination to provide the originals, that claim would be emptied quite fast. As is, we got something even worse than the PSB drivers, a bunch of binaries without even the source for the 2D driver, and Intel removed the source they had already made available on the Meego repositories. I think I might still have it around, so if I have time I'll check if it builds and if we can separate the open from the closed source in the EMGD driver.

Imagination can claim whatever they want, their stuff is not open-source so in any case, whatever Intel does, they cannot do anything significantly better than we have right now...

Please notice that Imagination does not even ship their SDK for intel, only a 'compatibility layer' for Intel. That means Intel must have some deal with Imagination because they do ship the closed source libraries in their driver (and with the SDK from ARM, we have the headers and stuff to actually use them too, but it does not originate from Imagination, altough it's their chip)

-Yves