Garlach44, a new pro-Trisquel company
I noticed Garlach44, an Italian company, offers a selection of laptops preinstalled with Trisquel GNU/Linux.
http://www.garlach44.eu/
http://www.garlach44.eu/en/for-free-software-users
:D It was just today that I found out about them from the gnu.org page. They do seem quite cool in fact! Would be awesome if they would donate some small part of their sells with Trisquel pre-installed to this project.
Does the AMD Radeon HD 6310 Graphics (used in Cork 13.3" AMD E-350t) support 3D acceleration in Trisquel?
Not on free distributions like Trisquel.
And the Nvidia GeForce GT540M? Does it support the 3D acceleration?
The free driver for NVidia says "Some 3D acceleration exists" for some chipsets. This was a reverse engineered driver from my understanding. Not ideal. The 3d acceleration is severely limited or non-existent. There is also non-free firmware that may be required in some (many/all?) instances.
It does not appear that there is ANY 3d support for the GeForce GT540M. I don't even know if it has 2d support. Which means you could be stuck with VESA. If I'm mistaken feel free to correct me. We won't work with NVidia/ATI chipsets because these companies refuse to work with the free software community. I am looking at one of the free reverse engineered NVidia drivers that I believe would be correct for this card. For at least some (maybe all?) of the NVidia chipsets that this free driver supports a non-free firmware component is required.
AMD has been better at supporting the free driver project. Don't let that fool you though. There is still no support for free distributions as there is a component that is non-free (firmware I imagine).
I don't want to put any one down although it is extremely frustrating how many people are attempting to do good in the community that don't understand the problems or have the ability to fix them.
I am going to point out that there are other companies where you can get a free distribution like Trisquel or GNewSense:
http://www.gnu.org/links/companies.html
Now... for the rant and why I dislike these companies (except the one I head, http://libre.thinkpenguin.com/).
These companies don't understand what they are doing is ugly. Very ugly. They are shipping hardware that doesn't work with Trisquel or GNewSense in many cases. They are shipping systems that will never work properly with any distribution. Free or non-free.
But... lets step back a second and ask some questions about why these companies should be avoided:
Will these computers work with free distributions like Trisquel? Only partially. If you don't care about your wireless working or your graphics working you will probably be OK! It isn't the only thing to worry about... but... these are two issues which frequently come up related to non-free drivers/firmware. I bet everybody cares if they can't access the Internet.
Some of the companies selling laptops with free software operating systems are shipping with wifi cards that don't work with free software operating systems like Trisquel (using Intel cards for instance). It gets worse though. You can't replace the wifi card in some of these laptops with one that does because the laptops include digital restrictions that are locking out other wifi cards. These are companies the FSF is linking to!
Do these computers work with non-free distributions? To a degree. They will work so long as the chipset manufactures continue to release updated drivers/firmware for the distributions. In many cases it may be necessary to manually install non-free firmware due to licensing restrictions/and or a driver each time the system is upgraded. This is frequently a complicated procedure that requires significant development experience. Particularly in situations where nobody has produced an easy to install package. The community in many cases can't support you. Neither the open source community which may support non-free software to a degree or the free software community. If you are manually installing non-free drivers, firmware, or other software you don't understand what free software is or why you should be using it.
Will these laptops/systems continue to work with the non-free distributions at least? Probably not. We have seen many chipsets (chipsets are things like wireless, video, sound, ethernet, etc) lose support due companies going bust or simply financial restrictions on the length of the support they can provide.
This does not strictly apply to laptops, desktops, netbooks, or tablets. This applies to everything. Printers, audio players, USB wifi cards, PCI wifi cards, PCIe wifi cards, and hundreds if not thousands of other types of devices. If you name a type of device chances are there is a device out there that does not work due to the chipset manufacturers refusal to cooperate with the GNU/Linux community.
There is a web site http://www.h-node.org/videocards/catalogue/en
where you can check various graphic cards with free gnu+linux
distributions like Trisquel.
I've said this so many time before and I'll state it again.
These lists don't work.
There is more to 3d acceleration than this.
Some NVidia cards have the bare minimum of support for 3d. It is the result of reverse engineering and not very good.
When I go and look at that list I have no idea what I would be getting based on comments like "it works" or "it doesn't work".
It is really quite misleading and only somewhat useful in limited circumstances.
Chris, I'm sure the h-node guys would be happy to hear improvement suggestions. Perhaps some tests to poke at different 3d features? Such could be included in the h-client currently in development.
h-node might be of limited use but isn't it the best thing we have still?
I did say it has limited value although I think it is generally misleading the users whom are using it.
Some cards on h-node have in their description that some programs using
3d acceleration work, some don't. Maybe they should recommend testing
it with more programs. I don't see a good way to check this
automatically in h-client (e.g. it shouldn't hang the machine with buggy
drivers and shouldn't make artificial results).
There are very good reverse engineered drivers for some hardware, there
are manufacturer-supported drivers which are magic without any publicly
available documentation. I don't consider RE leading by itself to worse
results. We know much more about microcode used on Nvidia cards than
one used on Radeons, despite the Radeon drivers being supported by AMD
and some documentation being available.
There were chipsets reverse engineered in the past for things like wifi. These chipsets were actually easier to reverse engineer in many cases than newer chipsets. They didn't have binary blobs or similar and still took forever. I can only think of and recommend one very old chipset that was reverse engineered. Reverse engineering is not the answer.
We need to be working with chipset manufacturers on the release and support of free drivers/firmware. I'll probably be announcing a new (cards are still being manufactured with this chipset, although the chipset is not really that new) chipset with free software support in the near future. Provided everything goes well.
There were chipsets reverse engineered in the past for things like wifi.
These chipsets were actually easier to reverse engineer in many cases than
newer chipsets. They didn't have non-blobs and the time it took was
significant. I can only think of and recommend one very old chipset that was
reverse engineered.
Don't you have to specify what driver you used for the card when making an entry to h-node? It would be sufficient, if there was some place to look up which drivers are reverse engineered and which ones are "real" (based on original documentations or made by the manufacturer itself). Sometimes it's even clear from the driver name: "nouveau + libgl1-mesa-dri-_experimental_"
There is other documentation elsewhere. It's not a workable solution. The majority of people (those not using GNU/Linux today) won't go find a list, purchase three cards, and continue using GNU/Linux. If you understand the way these products are made and advertised you can't be sure if the product you are getting is the one on the list. Nor the level of support it may or may not have.
Well it helped me to find card which can run Compiz and game I like to play.
Some cards have more comments than just "It works" like "I played that game with this card and it was ok" It depends on person who wrote information.
However it would be nice if we know which card works the best with Trisquel(with 3D acceleration) so we can recommend that card to users.
Do we have some better test than glxgears ?
That's very easy. There are basically just the integrated cards from Intel. Intel does the drivers on their own and gives the source to the guys maintaining the XServer.
You could also go out to a store and purchase several cards. You may get lucky and find one that works too. Then return the rest. The problem is it isn't a good solution. Not that it has no value whatsoever.
For a company who is researching chipsets and cards these lists may help. For the average user they are misleading, difficult, and turning people off of free software. Users shouldn't have to do such extensive research and use 'trial and error' approaches to finding hardware.
It isn't the solution. We don't have the resources to maintain them either. I'm saying this from experience. You can go to any of the major distributions and look at similar types of lists. The lists contain mostly hardware that isn't available. The quality of the information isn't great either.
I can't tell you how many products we have looked at based on these types of lists and discovered the chipsets were not the ones we were expecting. These projects start off great. Lots of hardware. A semi-usable list. Then die a slow and painful death. Along the way they frustrate users to no end. Frequently a piece of hardware that works today won't work tomorrow. In the case of h-node that doesn't apply generally speaking- although it does to most other lists.
Most people have limited experience with any particular type of product and make bad assumptions thereafter based on this inadequate experience.
I'll give you an example. I buy a printer. It is a 'Eexmark'. The Eexmark works for me. What you don't know as a consumer is the printer you bought has a very large defect rate. You got lucky though and it lasted a while. You go to buy a second printer. You think Eexmark is great because it lasted you a while. You even recommend it to others. Had you bought 10 of these and 5 in 10 failed in 6 months you probably wouldn't think this though. In fact it turns out Eexmark only sells defective printers for the most part. You see they "refurbish" returned printers. In practice what they actually do is plug the printer in, turn it on, and print a test page. Meanwhile every printer has had an issue. Some may not print a single page, most do. The problems come later.