It's not only NVIDIA with Linux problems that cause upstream developers to publicly bash companies, but AMD has come under scrutiny too. The developers of the popular cross-platform XBMC multimedia project shared a little story about enthusiasm, hope, and disappointment. In this guest posting by Peter Fruhberger on Phoronix, XvBA is what is principally talked about, which is AMD's lead choice for video acceleration when using their proprietary Catalyst driver. Unfortunately the XBMC developers aren't too happy about the state of video acceleration using AMD's Catalyst driver for Radeon graphics hardware, hence why they have reached out to Phoronix with this rather lengthy public message. Whether AMD even cares about Linux users and when XvBA will support missing functionality are among their open questions for AMD.
After sharing the results last week of an optimized open-source Radeon driver trying to compete with AMD's Catalyst driver, it is time to turn the tables. In this article is a look at the latest open-source Nouveau driver code compared to NVIDIA's official closed-source Linux driver across a few generations of GPUs.
There's been a number of recent open-source driver improvements that have come about for modern ATI/AMD Radeon graphics cards under Linux, but not all of these features have yet to be merged or enabled by default (e.g. 2D color tiling, PCI Express 2.0, and HyperZ). With some basic tweaks, can the open-source Radeon Gallium3D driver now compete with AMD's proprietary Catalyst Linux driver when it comes to OpenGL performance? Let's see.
At the request of Phoronix readers curious about the NVIDIA VDPAU performance between different GeForce graphics cards, here are some results.
The latest open-source Radeon Linux driver code for supporting the Radeon X1000 (R500) series hardware has basically reached a point of competitiveness with the legacy Catalyst Linux driver that once supported this hardware. In some cases the open-source Gallium3D driver is now faster than what Catalyst once ran at while in other select OpenGL workloads there is still a proprietary imbalance.
The Linux 3.5 kernel is capable of delivering some massive performance gains for some of the more recent generations of ATI/AMD Radeon graphics processors. Here's some benchmarks showing the hefty performance gains found when using the latest kernel that is still being developed.
Up as an extra article from Munich is a comparison of Intel DRM drivers on recent Linux kernel releases while using Core i7 Ivy Bridge hardware. With the Linux 3.4 kernel there are definite Intel Linux graphics performance gains.
Now having compared the graphics driver performance between Microsoft Windows 7 and Ubuntu 12.04 Linux for the NVIDIA driver with the GeForce GTX 680 and the multi-platform Intel performance for Sandy Bridge and Ivy Bridge, here's a look at the AMD Catalyst driver performance with the Radeon HD 7950 graphics card when running between Windows 7 and Ubuntu Linux.
Up for publishing today is a multi-card multi-driver comparison spanning several generations of discrete ATI/AMD Radeon graphics cards looking at the Radeon OpenGL performance as found by default in Ubuntu 12.04, as found when updating to the latest "Git" development code, when tweaking the latest development code for maximum performance, and finally when using the proprietary AMD Catalyst Linux display driver.
As the Intel Ivy Bridge benchmarks being delivered on Phoronix and coming up in the coming days are frequently using the latest Intel Linux graphics development stack, for those curious here is a comparison between the stock Ubuntu 12.04 packages and when running the latest Linux kernel / Mesa / libdrm / xf86-video-intel Git DDX.
Going back to last year there's been the "Glamor Acceleration" project out of Intel to accelerate the 2D operations within X using OpenGL on Mesa. This is similar to the Xorg state tracker approach and while it's not yet enabled by default, Intel OTC developers have been making much progress in recent months. In this article is a look at the recent Glamor update while comparing it to the stock Intel UXA acceleration as well as to the other experimental acceleration option: Intel SNA.
Last week the R600 LLVM compiler was hooked up for AMD's open-source Gallium3D driver. This LLVM shader compiler is important particularly for OpenCL enablement within the open-source Radeon Linux graphics driver, but still there is some ways to go before that code is ready for production use.
Thanks to clean-room reverse-engineering, it is already possible to run the NVIDIA GeForce GTX 680 "Kepler" graphics card on a fully open-source graphics driver complete with OpenGL acceleration. Here are the first benchmarks of this work-in-progress, community-created open-source GeForce 600 series graphics driver.
Committed to the Mesa and libdrm Git repositories last week for Nouveau, the open-source NVIDIA Linux driver, was the "major libdrm rewrite" designed to step-up this reverse-engineered driver. What impact did these invasive changes have on the Nouveau driver's performance? Here are benchmarks comparing before-and-after as well as how the Nouveau driver compares to the proprietary NVIDIA Linux graphics driver.
In what will certainly be controversial and disappointing to some Radeon Linux desktop users, AMD will soon announce that they will effectively be discontinuing support for several Radeon product families from their proprietary Catalyst driver. After that point, for future Linux distribution updates, the open-source Radeon Linux driver will be your only option for accelerated graphics. This is likely happening with the Windows Catalyst driver too, but at least there they have a better-maintained legacy driver process.
It's been a few months since last running any AMD Fusion tests under Linux, so here's a look at the AMD A8-3870K "Llano" APU performance under both the latest Catalyst driver and the open-source Radeon Gallium3D stack with Ubuntu 12.04. Besides the open-source driver being handily beaten by the Catalyst binary driver, the power efficiency is also a disappointment.
There is another new open-source Linux graphics driver entering development and it has already showed signs of success with basic 2D acceleration working. This new open-source driver is for Qualcomm's Snapdragon / Adreno and who is leading the development of this driver is also quite interesting.
Nouveau, the reverse-engineered open-source graphics driver for NVIDIA's entire range of graphics processors, is reaching a stable state where it's exiting the "staging" area of the Linux 3.4 kernel and being considered part of the standard, stable kernel configuration. How though is the Nouveau driver working out compared to NVIDIA's official, closed-source Linux graphics driver? Here are some new benchmarks from ten different graphics cards and other information on the state of Nouveau.
Following recent advancements in the open-source Radeon Linux driver like 2D color tiling support, I've carried out some new benchmarks of the open-source Radeon Gallium3D driver compared to AMD's official Catalyst driver. This time around, the open-source driver is seeing tests against AMD's binary blob when various performance-optimizing tweaks are enabled to see where the performance stands today.
In continuation of the Using The New Radeon Gallium3D 2D Color Tiling article from January, here's updated benchmarks of the latest Radeon Linux driver code with this performance-boosting feature enabled.
Recently I showed off some Intel DRM benchmarks from the Linux 3.3 kernel along with what will be in the Linux 3.4 kernel via drm-next. These results indicated some performance improvements on the side of Intel Sandy Bridge hardware, but how do the new kernel and Linux 3.4 impact the discrete Radeon graphics? Here's some benchmarks covering that side of the graphics table.
With the release of the Linux 3.3 kernel being imminent and the Linux 3.4 kernel drm-next already offering lots of changes, here are some Intel Sandy Bridge benchmarks comparing the Linux 3.2 kernel to a near-final Linux 3.3 kernel and then the drm-next kernel that's largely a 3.3 kernel but with the DRM driver code that will work its way into Linux 3.4.
After first being introduced on Windows years ago, and then FreeBSD and ReactOS support added last year, this week finally marked the release of TitaniumGL for Linux. TitaniumGL is self-described as a "freeware driver architecture" and carries a goal to support OpenGL on graphics cards with broken, bad, or missing OpenGL hardware drivers. Here are some benchmarks of TitaniumGL compared to NVIDIA's binary GPU driver and the Mesa/Gallium3D LLVMpipe software rasterizer.
Earlier this week I shared a pleasant surprise in Mesa 8.1 Radeon Gallium3D with some significant performance improvements to be found in the current Mesa Git code-base for the "R600g" driver in some OpenGL games. In this article is a more diverse look at the current state of Mesa 8.1 development for R600 Gallium3D and comparative benchmarks from every major release going back to Mesa 7.10.
Recently there has been a lot of talk about Hierarchical-Z/HyperZ support for R600g since its set to provide a measurable performance benefit the open-source Radeon driver for modern GPUs. This support has still not been mainlined and there are still a few issues to work out, but at least the support is stable for the older "R300g" driver and has been around for quite a while at this point. For those wondering what to expect from HiZ/HyperZ performance boosts, here is a preview.
Besides video decoding, re-clocking / performance improvements, OpenCL, and other areas, the Nouveau driver still has room to advance when it comes to 2D performance.
One of the less talked about features of Mesa 8.0 is its ability to handle MLAA, which is short for Morphological Anti-Aliasing. But how does MLAA on the open-source graphics drivers affect the OpenGL performance and is it worth it for boosting the image quality through this anti-aliasing technique? In this article are some benchmarks of MLAA under Mesa 8.0.
For the past few years VMware has been improving the graphics acceleration support that is available via their virtualization platform. VMware -- through their 2008 acquisition of Tungsten Graphics -- has effectively re-written their graphics driver for their virtual "SVGA II" GPU to take advantage of the Gallium3D driver architecture, a new acceleration architecture, and many other improvements. This work has finally come together and is now working rather nicely.
With the Mesa 8.0 release right around the corner, in recent weeks there have been a number of benchmarks on Phoronix looking at this latest open-source OpenGL library and its drivers, including Gallium3D. In this article though are new benchmarks from one of the areas not explored yet: the Intel Gallium3D driver performance.
While RC6 support remains off-by-default as Intel developers are faced by RC6-related bugs affecting a small minority of Sandy Bridge users, this power-savings feature is not limited to only Intel mobile graphics. As discovered at Phoronix, RC6 can manage to boost the graphics performance beyond just extending your battery life. The RC6 performance boost is also quite visible on Intel Sandy Bridge desktop hardware too.
716 display drivers articles published on Phoronix.