NVIDIA has put out two new proprietary Linux driver updates. One of the drivers is a pre-release in the 270.xx series that largely is after bug-fixing, but the second driver is more interesting as it's the first (development) NVIDIA Linux driver release in the 275.xx series. The NVIDIA 275.09 beta driver brings new features.
In continuation of the recent topic about NVIDIA Optimus coming unofficially to Linux, Red Hat's David Airlie has just pushed several patches into drm-next that deal with Optimus. These patches will be part of the DRM pull request to then go into the Linux 2.6.40 kernel once its merge window opens.
NVIDIA's Optimus multi-GPU technology now works under Linux. Well, at least for some notebooks, it's been hacked together by an open-source developer and in fact is working to use both Intel and NVIDIA graphics processors simultaneously with the respective drivers. This is the best Linux implementation we've seen yet with NVIDIA Corp still not announcing plans to officially support this technology under non-Microsoft operating systems.
Aside from political issues surrounding open vs. closed-source (graphics) drivers on Linux, the proprietary NVIDIA Linux driver is widely liked. The proprietary NVIDIA Linux driver is relatively bug/trouble-free, has a performance parity to the Windows driver, supports new hardware right away, and has a near feature parity to the Windows driver. There's not much more you could ask for from a closed-source driver, aside from a few missing features. One of the missing features that's been widely talked about as of late has been Optimus.
NVIDIA has updated its legacy binary Linux display driver. The NVIDIA 173.14.30 is this new driver release and it simply adds support for X.Org Server 1.10 and compatibility for the newest Linux kernel releases (up through Linux 2.6.38). That's it.
The NVIDIA crew working on their proprietary Linux driver have just pre-released a new build, NVIDIA 270.41.03. This Linux driver update mainly adds support for a number of new GeForce / Quadro GPUs.
A call for testing has been issued on NVIDIA's binary Linux display driver for the upcoming Ubuntu 11.04 release.
One of the features that's supported by NVIDIA's binary Linux driver that is not supported -- nor has even been attempted -- by the community Nouveau project or any other open-source project is for 3D Vision / 3D Vision Pro. 3D Vision is NVIDIA's technology that combines their consumer and workstation GPUs with specialized glasses and capable displays/projectors to provide a realistic 3D experience. 3D Vision Pro is effectively the same but with a focus upon the professional/enterprise markets by creating an immerse experience in Autodesk, Maya, and other costly applications.
Due to RandR 1.4 being pulled from X.Org Server 1.10, the video driver ABI had to be bumped to again, and this was at the last possible minute with X.Org Server 1.10 being released just days later. For the open-source X.Org drivers this just means recompiling the driver for the latest binary interface. For the binary blobs, this means NVIDIA and AMD must put out new releases.
NVIDIA has announced the release of the CUDA 4.0 Tool-Kit this morning, which continues to be fully supported under Linux. NVIDIA's Compute Unified Device Architecture 4.0 focuses upon GPUDirect 2.0 Technology, Unified Virtual Addressing, and Thrust C++ Template Performance Primitive Libraries.
And so it begins. Less than 24 hours after presenting OpenBenchmarking.org on the weekend, dozens of results are beginning to pour in (the latest test results). In particular, there's already a few sets off results that may interest many Phoronix readers.
Last week when talking about NVIDIA looking to expand its Linux team (hire more engineers), I asked what else NVIDIA Linux customers wanted that already wasn't offered by the proprietary driver for Linux / BSD / Solaris operating systems. Aside from the obvious one, of many desktop users wanting NVIDIA to support some sort of an open-source strategy, other expressed views are listed below.
In early December I shared that NVIDIA was looking to expand its Linux driver team after Andy Ritger mentioned they were looking to take on new engineers. It appears they're still looking to expand with an updated job posting yesterday seeking a Linux graphics software engineer.
This morning I talked about a stable NVIDIA Linux driver update (v260.19.36) and that a 270.xx beta driver would be imminent. It turns out, however, that the NVIDIA 270.18 Beta driver is already publicly available. It can be tested for Linux x86/x86_64 with a couple of new features to this proprietary graphics driver.
On Friday the NVIDIA engineers working on their proprietary Linux driver put out a new driver update. This new driver update is marked NVIDIA 260.19.36 Certified, but it brings just three official changes.
NVIDIA has announced from the Consumer Electronics Show in Las Vegas that it's working to deliver ARM CPUs for a range of devices from PCs to super-computers. NVIDIA plans to build high-performance ARM CPUs for a range of devices, including servers and workstations. Internally this is being worked on at NVIDIA under the Project Denver codename.
Two weeks after NVIDIA had put out their unannounced 260.19.21 Linux driver, they have returned to the web and have officially released the NVIDIA 260.19.26 graphics driver for Linux x86/x86_64 along with Solaris and FreeBSD operating systems.
Over the week I mentioned that NVIDIA may be working on a new driver architecture for its unified, proprietary GPU driver on Windows/Linux/Solaris/BSD platforms. This information was learned when one of NVIDIA's Linux engineers was engaging in a technical debate over their fence synchronization patches for the X.Org Server. Andy Ritger who heads the NVIDIA Unix Graphics Driver team, has provided us some brief comments with regards to their ongoing architectural work.
When going over the mailing list messages from the past few days regarding concerns over NVIDIA's fence sync patches for X.Org Server 1.10, one of the statements by NVIDIA's James Jones indicates that they are working on a new driver architecture. What though could this new driver architecture hold in store?
As reported this morning, RandR 1.4 is now ready for X.Org Server 1.10 after this next xorg-server release's merge window was kept open to allow this work to be finished and land in the Git tree. RandR 1.4 brings per-CRTC pixmaps, sprite transforms, and a new RandR request that will hopefully allow NVIDIA's binary driver to finally support RandR 1.2+ features. While the merge window is kept open for a short period of time, NVIDIA has been trying to put in their Fence Sync support for the X Server. While it looks like these patches will still be accepted, some objections and questions have arose over this open-source contribution by NVIDIA.
X.Org Server 1.10 was just looking to be a big bug-fix release to the X.Org Server with no major features being introduced, up until the merge window was about to be closed. Then last night it was proposed by Keith Packard, the xorg-server 1.10 release manager, to keep it open a few extra days so that he could finally merge the per-CRTC pixmap support. This work alone is nice and is long awaited, but now NVIDIA's James Jones is calling for pulling another feature that's had code available for months: X Synchronization Fences.
To those that missed it in our Phoronix Forums where it was discovered, NVIDIA two days ago uploaded a new binary Linux (x86/x86_64) driver to their FTP server. This NVIDIA graphics driver is marked as the 260.19.26 beta, but they have yet to officially announce this new release or even provide a change-log.
While NVIDIA should soon be releasing a new Linux graphics driver beta, for those of you interested in NVIDIA's Compute Unified Device Architecture (CUDA) rather than -- or as a complement to -- OpenCL, there is a new tool-kit release. CUDA 3.2 is now available this week. CUDA 3.2 brings a number of new features to the NVIDIA GPGPU table.
Earlier this month NVIDIA rolled out the GeForce GTX 580 graphics card as their fastest GPU to date with 512 CUDA cores, a 772MHz core clock, 1544MHz processor clock, 1536MB of 2GHz GDDR5 memory, and support for three-way SLI. The GeForce GTX 580 with its GF110 core is based upon a refined version of the Fermi architecture and is certainly a step-up from the GeForce GTX 480 that launched just earlier this year. For those curious how this NVIDIA graphics card performs under Linux, here's the first benchmark and it's compared to the Windows driver performance too.
It was just one week ago that NVIDIA released a stable Linux driver update, but today for those wishing to live on the bleeding edge of NVIDIA's proprietary Linux driver development, the first beta release in the 260.xx series is now available for testing. The NVIDIA 260.19.04 Linux driver brings a lot to the table.
Over the weekend there was a new Linux binary driver release from NVIDIA that was the 256.52 driver in a pre-release state. It didn't deliver on OpenCL 1.1, Fermi Linux overclocking support, or any other radical features, but it did bring a handful of bug-fixes. Today this driver has been released as stable after being branded the NVIDIA 256.53 driver.
Just shy of a month ago was when NVIDIA last released a proprietary Linux driver, at which point they also released a second driver that was their OpenGL 4.1 preview driver. This Saturday though NVIDIA has provided a new driver release, which is tagged as the 256.52 pre-release. This new Linux driver release isn't overly exciting, but it does carry some prominent fixes that will please some NVIDIA customers.
Next week we will finally be able to deliver performance numbers for the NVIDIA GeForce GTX 400 "Fermi" graphics cards as we take a look at the GeForce GTX 460 that was kindly sent over by NVIDIA. Overall it's a very interesting card and great performer on Linux, but if you're already a Fermi owner (ideally after buying the hardware with our shopping links) and have been searching the Internet like we had done wondering why CoolBits isn't working on Fermi hardware, well, we now have the answer.
Back in 2006 a start-up company known as AGEIA launched the PhysX PPU, the first Physics Processing Unit (PPU) for offloading physics calculations in games and applications that utilize the PhysX API onto this discrete processor for boosting overall system performance.
Back in June there were patches published by NVIDIA for X Synchronization Fences after it was in planning since before last year's X Developers' Summit. These synchronization fences were designed to "synchronize X rendering with direct rendering X clients such as OpenGL and vice versa. They are especially useful in synchronizing GL-based composite managers screen updates with Xs rendering." Today there are more patches out from NVIDIA this time touching xextproto to add support for the binary sync objects.
538 NVIDIA news articles published on Phoronix.