I'm not a Linux user mainly but familiar with Sandy Bridge integrated graphics. Tested this iGPU some weeks on Windows 7. Overall my tested games did run pretty good without glitches. Sometimes issues with 3D effects like HDR or Motion Blur but it happened rarely. Some of my other issues I had summarized in short:
- missing Vsync "Off" option in control panel
- Vsync option "On" available but did not work at all
- Anisotropc filtering option in control panel not working as well in most of the games when set to 16x, 8x etc. (placebo option similar to Vsync)
- Sandy Bridge iGPU supports 2x and 4xMSAA, what about games without MSAA option in game?, there is no MSAA option available in Intel Control panel
- 23.976 fps video playback bug
I'd like to begin by saying that I find what you (Intel driver team) are doing is just amazing, for maybe 15 years I have been dreaming for hardware manufacturers to seriously invest in open support for their products, and after all these years you are one of the few who seem to really want to do things right. Graphics drivers being as primordial as thy are (I can work without printer, webcam or wifi) I think your work is very important and you seem to be doing a great effort, with very good results. I have bought Intel graphics systems for years, and even if it has not been smooth riding all the way, it has always been far from the terrible headaches I had with the two other vendors, and in recent years the experience has become much smoother - recently I have had hardly any issues, I even find the situation better now than I used to have in windows. It just tends to work, great !
I recently bought a laptop however, and by default the backlight just wouldn't light. I know this is peripheral to your main work, and I managed to find a workaround (after days of effort), but now I must manually switch the light on with a key combination. Maybe I shouldn't complain too much, as it can be made to work, but I consider unsatisfactory any system for which the screen does not work out of the box, and really on a laptop I shouldn't have to switch the screen on independently from the machine. I filed a bug, and after a little complaining I got decent attention and I expect that this is fixable. This issue is present in many machines across vendors, and has been present for over a year - hundreds of users are complaining on forums across distributions, and it is not easy for a non techie to work around.
Now to my question : could you not buy a few tens of common machines, with diverse hardware configs, then have an intern test your drivers on each before any release ?
I noticed you say "I don't have any i7 around" for testing, if you don't have one around, I do hope you have access to one. Looking at the problems in bugzilla, it would seem to me that if you installed your drivers on a range of hardware you would come across many of these issues yourself. With testing on more configs, couldn't you avoid having users trip over the bugs ? Wouldn't it be easier to fix problems if you yourself had the offending system in front of you, rather than diagnosing via bugzilla ?
1) New generation is more energy efficient if you compare performance per watt with vintage hardware. You will spend 20$ for new hardware but how much you will save on electric bill every month?
FYI: energy cost to much not everywhere. In my case I can doesn't care about such thing at all. I just want to keep old hardware up and running. And, you know, I am not only one who want that.
Originally Posted by zbiggy
2) This one is OK. But today netbook has power of mainstream PC box from year 2000. Smaller PC = less waste.
And so what? So what if PC box from 2000 year up running and work fine for their user? They should replace it just because there is something new?
Look at post christobzr above - new machine not always work better.
"We do not develop those solutions at Intel, those are being carried out by the vendors (Nvidia and AMD). So there is not much I can comment about that. However, by default, machines boot with integrated cards by default for power saving reasons."
I call BS and that this is a cop-out. Please do not just come here to astroturf! (There is enough of that from ATI.... Wait till our next drivers? Years pass.) Intel does develop this in collaboration with the other vendors and ditto the bios. More so in the latter case.
As Intel well knows, Linux users want the option to turn off the lower specd onboard graphics and use the real deal. The latter has been the norm for some time now. Pretending ignorance (for well over a year) and asking "Intel Wants YOUR Linux Questions, Feedback," is a sham.
So then, please put what was always there, there is a good Intel! ;-)
To be fair, i810-like series of chipsets went out in production several years ago, are not manufactured for several years ago, and even the longest possibly warranty for them has already expired years ago as well. But they still work, just do not support the newest features.
This is not true, e.g. 845G doesn't work for years. GPU crashes. No 3D, no 2D accel., no XVideo. Only VESA, and VT switching under it leaves black screen.
Originally Posted by eugeni_dodonov
Just like with windows drivers, you cannot expect to install Nvidia drivers for latest series of their graphics cards, and expect it to power up Riva TNT2 card and allow you to run Crysis. But if you do install the supported drivers (which you can find in form of older kernel releases + older mesa releases + older xf86-video-intel), they will work .
The great thing of the OSS drivers should be good and still working legacy HW with new OS and technologies. This is an argument, like, well why don't you install Windows 98 beacuse we won't port driver to XP. Nobody expect Crysis on 2nd generation Intel. But working 2D and XVideo acceleration at least, yes.
Originally Posted by eugeni_dodonov
I am not a commercial nor marketing person, so I cannot give any dates .
What I can say is that the hardware is supported as long as it is being manufactured + as long it is being warrantied + as long as we have customers using it in large enough scale to justify allocating resources for its support. After that time, it can be supported if there are people interested in such support.
2nd generation Intel is still working all around the world. Intel flooded the market with it in early 2000s and the machines are still alive and in use.
Originally Posted by eugeni_dodonov
With our drivers, specially considering Sandy Bridge architecture, we have a huge bonus though - our specifications are open! So if necessary, the community is able to implement new features or support new stuff even without our direct intervention. Lots of new features were developed for the drivers without our team - the community did it. This is what the true power of open-source is IMHO.
I think this is the crucial issue. The problem with 2nd generation is that, AFAIK, detailed documentation wasn't released and because of undocumented quirks in HW, after introduction of new memory management for drivers in X.org, support blowed-up. I think Daniel Vetter is trying to fix it for a long time, but probably because he doesn't have a deeper insight into HW quirks, the satisfactory solution is still not found.
I know that it's probably a lawyer issue, but it's a 10 years old hardware. What harm is in releasing specs? Or at least someone from Intel with the deeper knowledge of the achitecture quirks give some hints to the community to fix those problems?
Intel GMA 4500 is horribly broken in Mesa 8 devel, tested Mesa 8 rc1 and and so forth.
Only things work is glxgears and desktop 3d effect everything else like games getting opcode fail or segfault like below:
X Error of failed request: BadRequest (invalid request code or no such operation)
Major opcode of failed request: 34 (X_UngrabKey)
Serial number of failed request: 44
Current serial number in output stream: 44
Tested on latest Ubuntu 11.10
Xorg updated to X.Org X Server 220.127.116.112
Linux kernel 3.2.0-030200-generic
I have an Atom N450 in my netbook. According to your website it should support OpenGL 2.0 under linux.
My glxinfo output however reports OpenGL 1.4 (that one was made with Mesa 7.11, but switching to xorg-edgers ppa which holds 8.0-devel doesn't change anything).
The hardware is capable of OpenGL 2.0 however, since switching over to the community Gallium3D driver (in mesa-experimental) produces this glxinfo.