It has been two years since the ATI Radeon HD 4800 (RV770) series launched so we have gone back since that monumental hardware launch and have re-tested each Catalyst driver release since then to see how the performance has changed for the ATI Radeon HD 4850 graphics card. The Catalyst driver has certainly matured over the course of two years in speeding up the OpenGL performance with this hardware along with bringing new features to their proprietary driver, but it is not exactly smooth sailing.
Earlier this week AMD released the Catalyst 10.6 driver that on the Linux side of the table had finally made use by default of their new 2D acceleration architecture, offered official support for Red Hat Enterprise Linux 5.5, and formalized their OpenGL 3.3/4.0 support. Since the release of the Catalyst 10.6 Linux driver, we have been running a new set of tests on their new ATI 2D acceleration architecture, but the results are not what you may expect when compared to the open-source ATI Linux driver.
A number of weeks back a set of benchmarks were published showing even the latest open-source ATI 3D driver is still no match to an old Catalyst driver and that even Gallium3D lags behind the Catalyst driver for those interested in OpenGL gaming. However, in other areas the open-source ATI driver stack is beginning to win by measurable amounts.
Last week prior to heading over to Germany for LinuxTag, I had ran a new set of ATI R500 Gallium3D benchmarks with an ATI Radeon X1950PRO graphics card and comparing the latest Mesa/Gallium3D graphics driver performance in the Mesa 7.9-devel Git code with both the Gallium3D and classic Mesa DRI drivers to the older Mesa stack found in Ubuntu 10.04 LTS. The ATI "R300g" driver as its known continues to advance, and over the past week this driver has pushed forward even more. Here is another set of ATI Gallium3D tests.
There was a talk last week at LinuxTag in Berlin by Egbert Eich about kernel mode-setting and the DRM (Direct Rendering Manager) graphics stack on Linux. Egbert is, of course, a long-time X developer and openSUSE developer at Novell who was one of the masterminds behind the RadeonHD graphics driver and has worked on various pieces of X over the years. In Egbert's brief KMS talk he briefly covered the history of the Linux graphics stack, the user and kernel-space APIs for DRM mode-setting, and related topics. For those that missed out on his talk, below are his slides.
The past several months have been very exciting in the world of Gallium3D, the new graphics driver architecture for Linux and other operating systems that has been in development for years. This year we have witnessed the emergence of LLVMpipe to accelerate OpenGL commands on the CPU, Nouveau's Gallium3D driver starting to work well, and many other advancements. Over the past few months we have also been pleased with how the "R300g" driver has taken shape with this Gallium3D driver for ATI Radeon R300/400/500 series hardware (up through the Radeon X1000 series) stabilizing, performing well, and advancing beyond the classic Mesa 3D R300 driver. Today we have a fresh set of benchmarks looking at this ATI Gallium3D driver that soon will become the default.
Last week NVIDIA released their first 256.xx proprietary beta Linux display driver that brought many VDPAU improvements, installer improvements, support for new GLX extensions, various bug-fixes, and other enhancements. However, some user reports have shown the 256.xx driver is actually slower than NVIDIA's current pre-200.xx series drivers and so we have carried out a set of tests to see what things are looking like from within our labs. Our preliminary tests do indeed illustrate a drop in performance when upgrading to this new driver.
Yesterday we reported on VIA's Linux dreams not materializing with their GEM/TTM memory management support still missing even though we are half-way into 2010 -- more than two years after VIA announced its most recent open-source initiative. It turns out, however, for what VIA views as its memory management work is actually done. VIA has inconspicuously handed over some of its code to the OpenChrome developers in order to create a new driver that has been dubbed the "openvia" driver. VIA has supposedly provided the source-code to an X driver plus TTM/GEM DRM, but this new project largely remains a hidden mystery.
Last week we reported that the open-source ATI Linux driver had picked up improved power management in the form of dynamic power management and power management profiles that can be defined by the end-user. With the ATI Linux power management finally coming to fruition within the Linux kernel for its kernel mode-setting / DRM driver, we have decided to take a close look at how this power management support is working in the real world.
The software rasterizer used in Mesa that allows for software acceleration of OpenGL on the CPU without any assistance from the graphics processor has largely been useless. Even with a modern-day, multi-core processor, the performance of Mesa's software rasterizer has been abysmal. The performance of Mesa classic DRI drivers have traditionally been poor anyways compared to the high-performance, proprietary NVIDIA/ATI graphics drivers, but when dealing with just the software rasterizer there really aren't any games or applications that run well. Fortunately, software acceleration on Gallium3D is very much a different story thanks to LLVM.
We have already published a look at the Fedora 13 Beta, delivered ATI Radeon benchmarks atop Fedora 13 Beta, and have other articles on the way covering this new Fedora release, while in this article we are investigating Nouveau's power performance using this newest Fedora release. If you are a mobile user planning to use the Nouveau stack right now, or you care the least bit about energy savings with your desktop, its power consumption alone may rule this open-source driver out as even a current possibility.
With the release of Fedora 13 Beta earlier this week we have been testing out this Red Hat update on a few of our test systems. One area of interest to us has been to see how the open-source graphics are performing with Fedora 13, since after all Red Hat is known to always ship the very latest DRM/Mesa/DDX bits in Fedora due to all of their upstream involvement and this week is also the Fedora 13 Graphics Test Week. We already looked at the direction of Intel graphics with Fedora 13, so our next target was testing out the open-source ATI graphics with this Linux desktop release that is codenamed Goddard. In this article, we have ATI R500 tests using their open-source driver stack as we test out the OpenGL performance and the power consumption, compared to Fedora 12.
Yesterday we delivered benchmarks showing how the open-source ATI Radeon graphics driver stack in Ubuntu 10.04 is comparing to older releases of the proprietary ATI Catalyst Linux driver. Sadly, the latest open-source ATI driver still is no match even for a two or four-year-old proprietary driver from ATI/AMD, but that is with the classic Mesa DRI driver. To yesterday's results we have now added in our results from ATI's Gallium3D (R300g) driver using a Mesa 7.9-devel Git snapshot from yesterday to see how this runs against the older Catalyst drivers.
We are in the process of conducting a set of tests looking at how the performance of Ubuntu Linux has evolved through their Long-Term Support (LTS) releases beginning with their first 6.06 "Dapper Drake" version followed by Ubuntu 8.04 "Hardy Heron", and then the Ubuntu 10.04 "Lucid Lynx" release that will be released by month's end. These benchmarks will look at how the performance of Ubuntu Linux has changed over the past four years, but first we deviated from our original plans to get a look at how the current open-source ATI R500 graphics driver in Ubuntu 10.04 provided by the Mesa stack performs against older proprietary ATI Catalyst drivers.
Over the past two weeks, we have published a variety of articles looking at different aspects of the open-source Linux graphics stack. These articles range from comparing the Gallium3D and classic Mesa performance to comparing the kernel mode-setting and user-space mode-setting performance. Today we are continuing with this interesting Linux graphics coverage by publishing benchmarks comparing the performance of the Radeon Mesa DRI graphics driver to AMD's Catalyst 10.4 proprietary driver. Is the open-source driver finally catching up to AMD's highly optimized driver? Continue reading to find out.
NVIDIA's open-source Linux efforts as it concerns their GPU support have historically been minimal. The xf86-video-nv driver has been around that provides very basic 2D acceleration and a crippled set of features besides that (no proper RandR 1.2/1.3, KMS, power management, etc) while the code has also been obfuscated to try to protect their intellectual property. However, NVIDIA has decided to deprecate this open-source driver of theirs. No, NVIDIA is not working on a new driver. No, NVIDIA is not going to support the Nouveau project. Instead, NVIDIA now just recommends its users use the X.Org VESA driver to get to NVIDIA.com when installing Linux so they can install their proprietary driver.
Last week we published benchmarks looking at the ATI Radeon KMS vs. UMS performance and found the user-space mode-setting support with the ATI driver (that is also limited to using DRI1 with these older code-paths) to perform significantly faster than the newer kernel mode-setting routes in most instances. To see how the performance difference is on the Intel side between the kernel mode-setting and user-space mode-setting implementations we ran a set of benchmarks on this side as well using Ubuntu 10.04.
With Mesa 7.8 arriving this month, we took the time to benchmark a few recent releases of the Mesa 3D stack with the Radeon DRI driver to see how the OpenGL performance has changed -- if at all -- over the past few months. In this article are our R500 Mesa benchmarks from the Mesa 7.6, 7.7, 7.8-rc1, and 7.9-devel releases.
Earlier this month AMD rolled out a new workstation graphics card driver, which is effectively the same Catalyst driver used by the consumer-oriented Radeon graphics cards but with greater testing and certification for the ATI workstation offerings. The press release announcing this new driver was titled "Application Performance Increases By Up To 20 Percent with Latest ATI FirePro Graphics Driver," so we decided to see if this proprietary driver really lives up to its claims under Linux.
Gallium3D, the graphics driver architecture started by Tungsten Graphics to overhaul the hardware driver support in Mesa, has been around for a few years but it is finally getting close to appearing on more desktop systems. Now that the Nouveau DRM code is in the mainline Linux kernel and its main 3D driver is Gallium3D-based, we will hopefully be seeing that adopted by more distributions soon -- it's already being flipped on with Fedora 13. On the ATI side the "r300g" Gallium3D driver that provides Gallium3D support for the R300-R500 (up through the Radeon X1000 series) is also being battered into surprisingly good shape. To see where the Radeon Gallium3D support is at for these older ATI graphics cards we have run a set of tests comparing the OpenGL performance under the latest Mesa 7.9-devel code with the Gallium3D driver to running the classic Mesa DRI driver.
Earlier this week we published comparative benchmarks of Mandriva, PCLinuxOS, Ubuntu, and openSUSE. In the discussion that followed, a number of people requested a set of tests that compare the performance of the ATI Radeon Linux graphics driver stack with kernel mode-setting (KMS) vs. user-space mode-setting (UMS), so today we have such results to deliver.
As we shared a few days ago, Fedora 13 will provide OpenGL acceleration support for NVIDIA graphics cards via the Nouveau driver when installing the Mesa DRI experimental drivers package. There is finally 3D acceleration for NVIDIA graphics cards using an open-source driver on Linux without having to depend upon NVIDIA's official binary driver. What makes this open-source 3D support for NVIDIA GPUs even more interesting is that it is atop the Gallium3D driver architecture rather than classic Mesa. With that said, we are providing early benchmarks of the Nouveau Gallium3D driver in Fedora 13 with two GeForce graphics cards as we compare the performance to NVIDIA's official Linux driver.
As we alluded to last week, we have been in the process of benchmarking many Radeon HD 2000/3000/4000 series graphics cards using the open-source ATI Linux graphics stack with the Mesa R600/700 DRI driver. We have now carried out our first batch of R600/700 3D tests using this constantly evolving open-source driver to provide OpenGL acceleration and here are the results.
Yesterday Luc Verhaegen gave a talk at FOSDEM on reverse engineering a motherboard BIOS, but today we finally have X@FOSDEM for the last time. Luc has just begun his talk on unifying and simplifying the free software desktop's graphics driver stack. Here are his slides and we will be back with more updates and videos on Phoronix as the presentation progresses.
We know that NVIDIA's Video Decode and Presentation API for Unix (VDPAU) works very well for exposing PureVideo capabilities on Linux. We have benchmarked VDPAU and found it to perform very well in that under Linux it's possible to play HD videos with a $20 CPU and $30 GPU thanks to this video acceleration method. VDPAU is the best video acceleration / decoding API on Linux and is widely adopted by various multimedia applications, which is all in contrast to AMD's XvBA and their troubled implementation. But how does VDPAU work on mobile devices? With the ASUS Eee PC 1201N that is built on NVIDIA's ION platform we ran a new set of VDPAU video playback tests.
Compared to past years when recapping the AMD/ATI Linux advancements over the past calendar year, 2009 was not quite as exciting, which can be viewed as both good and bad for their Catalyst Linux driver. There were many advancements this year on AMD's open-source side, but in 2009 there wasn't as many milestones for their Catalyst driver like in the past with the introduction of CrossFire, OverDrive, same-day Linux support, the AMD Catalyst Control Center, and other new features. Here is our 2009 year in review look at AMD's advancements to their proprietary Catalyst Linux driver along with our annual benchmarks.
Nearly two years ago at the Linux Foundation Summit in Austin was VIA's most recent announcement about becoming serious with open-source support. This was not VIA's first time they claimed to back an open-source strategy, which led a number of open-source developers to immediately call VIA's open-source strategy a bluff. To date this still is mostly a bluff, but they have produced some fluff. In 2010 it looks like this will still be the case, but VIA hopes to produce some code by the second half of 2010. This code, however, will likely not appear in most Linux distributions until 2011.
Another annual tradition of ours besides running a Linux Graphics Survey is to provide a "year in review" analysis of the ATI and NVIDIA Linux drivers with their respective graphics driver releases from the past year in terms of both feature improvements and how their quantitative performance has changed. We have been doing these annual ATI and NVIDIA yearly reviews going back to 2005, but now it's time to share our thoughts and numbers for 2009. We are beginning with our NVIDIA Linux 2009 Year In Review.
Recently via email we were asked to run a comparison of the different anti-aliasing and image rendering options between the ATI/AMD and NVIDIA Linux drivers and hardware. Well, we have now run a few quantitative and qualitative tests at different anti-aliasing levels under Linux. For those that want to run the tests themselves with their own drivers and hardware, we also have provided instructions on how you can easily do so using the Phoronix Test Suite 2.4 "Lenvik" development build -- it is irresistibly easy.
For the month of November we ran the 2009 Linux Graphics Survey, which is a survey in regards to X.Org and the Linux graphics stack that we have been hosting annually for the past three years. This year there was 13,836 results submitted and we have now had the time to go over these results and are publishing all of the numbers today.
815 display drivers articles published on Phoronix.