High Performance? Really? Did anybody compare OpenGL rendering performance of Intel proprietary driver and Intel open source driver?
Yes, Phoronix: Intel Graphics: Windows 7 & Windows 8 vs. Ubuntu 13.04 & Fedora 18. If you sum up and average the results on each test for "Ubuntu Fedora/13.04+git" and "Windows 8" then Windows 8 comes out at an average of 3.5% more fps overall, which is negligible. (Linux wins some tests, Windows 8 others, the biggest win for Linux is OpenArena 800x600, the biggest win for Win8 is Xonotic 1366x768).
Like I said earlier, it would be very interesting to see this "15-way" test re-done next month when Haswell is released.
I think you are wrong in your idea how amd makes desitions. Its not like they dont want to give the community a top notch opensource linux driver, and they want instead have a fast blob and dont care to much about the open driver.
I am pretty shure that its more like that: "we have a working nearly complete driver like the fglrx, we have few resources we cant hire 100 developers for linux-driver only, it will take some years to get the free driver top notch, so we should minimaly support the non-free driver, we should update it to new abis, and fix some of the biggest bugs, but other than that this driver is marked for dieing in the future"
I thought I've read AMD guys basically saying "the closed source driver is aimed at gamers/graphics professionals, the open driver is for everyone else" but maybe they didn't. The original proposal from Novell suggested that the driver was intended for basic desktop acceleration and Phoronix said the same, there was no suggestion that the closed driver was going to be only minimally supported: ATI/AMD's New Open-Source Strategy Explained
Originally Posted by Phoronix
The aim of this open-source driver is not to overtake the fglrx driver but rather is designed for those who just want a working desktop with 3D capabilities and basic video playback. This new driver is ideal for FOSS enthusiasts and those wishing to run the latest development kernels and versions of X.Org. The fglrx driver will continue full steam ahead with their monthly releases and will be for those who want a stable driver with top-notch performance, all of the bells and whistles, and avoid checking out the latest git code in order to get the latest fixes and features.
Originally Posted by blackiwid
But maybe I am totaly wrong here, and amd has 100 fglrx-linux-only developers and only 5 radeon developers or so???
That is an interesting question, I'd imagine that they are focusing development effort on the game console and Android platforms, so which driver do you think Sony wants for PS4? Which drivers will their 64-bit ARM ship with? A large and complex port of a closed source driver that started life on Windows, or a smaller and sleeker driver that lacks performance optimisation?
nvidia blob is a shit? please don't joke, if you are an amd fanboy and you love the shit performance on serious sam o team fortress or any other game that use opengl ( no shits like urbanterror, openarena etc), do what you want and purchase all card exceeding great amd. But today amd drivers are crap on gnome shell, are crap with wine, are crap with xvba , for example when you use mplayer vaapi, when you swith to full scren at windowed mode, you can see artifacts, are crap on more demanding games, are crap with compiz perfomance. If you want to use community driver ok, but I want the support of the company that I pay for their prodcuts like on windows, and not the support of the community.
As of April 26, 2013, Catalyst packages are no longer offered in the official repositories. In the past, Catalyst has been dropped from official Arch support because of dissatisfaction with the quality and speed of development. This time, it's the incompatibility with Xorg 1.14.
It's a shame that amd is so slooow to support the latest technologies on linux and force you to use ubuntu.Amd need to learn of the nvidia strategy on this market.
That statement is 6 years old, it was made shortly after amd bought ati, and they did 1-2 years after that made some bigger improvemnts since then nothing big happend anymore, maybe I am wrong I dont use the blob so its maybe wrong. but if much would have happend not everybody would still bitch about that driver ^^
So again, yes they give kind of support, because it takes long to make a good free driver especially with software patents and very few linux driver developers on amds site at all.
I think, they yes give "support" in that kind that they will recompile and make it work on newer xservers, and they kind of bring because of the sharded code space of the windows driver some stuff that gets better there also to linux.
But I dont see that they improve fglrx faster than the radeon driver, I think the opposite is true, so they focus more on the free driver.
In the other news.... I burned my mobility X1900 today xD
I was just watching youtube html5 video on (older kernel 3.7/Mesa8 stack) in 720p using opensource driver. Obviously without any recently added acceleration (UVD1 is unsupported anyway).
Then I get kernel panic... On next reboot I have green squares everywhere xD. No mater what I do, the pattern changes on every resolution, but stays constant on each of them.
I am pretty sure I burned a memory chip.
At first, I thought its because of power management. That had probably contributed.. but then I read that X1xxx mobile cards were defect, like that of Nvidia G80 and basically need baking to resolder.. I probably sell this defect card.
So, right now I am looking for an MXM VGA... I am looking for an AMD MXM VGA, yes, I like this opensource driver very much now.
You guys should probably stop arguing and just buy cards you want to question (used) and do own tests... otherwise this won't end....
I managed to grab HD5850 used for 45$. You can too, you are not loosing much.
I'll see what I can do. Some of the Gimp 2.8 OpenCL filters already run on the r600g driver (you have to start gimp with GEGL_USE_OPENCL=yes to get it to use CL), but I haven't tested all of them yet... First, I have some repeatable (but not every time) GPU hangs that I need to nail down.