1. Computers
  2. Display Drivers
  3. Graphics Cards
  4. Memory
  5. Motherboards
  6. Processors
  7. Software
  8. Storage
  9. Operating Systems


Facebook RSS Twitter Twitter Google Plus


Phoronix Test Suite

OpenBenchmarking.org

NVIDIA AYiR 2006

Michael Larabel

Published on 5 December 2006
Written by Michael Larabel
Page 2 of 7 - Comment On This Article

After months of anticipation launched on October 13 was the first official Linux display driver in the 1.0-9XXX series. Previously, however, there was the 1.0-9625 beta display driver. In September we had seen an unprecedented move by NVIDIA where they had finally began offering public beta display drivers for its alternative OS users -- not only Linux but also FreeBSD and Solaris. Premiering with the 1.0-9626 display driver was support for the NVIDIA Quadro Plex VCS, GLX_EXT_texture_from_pixmap support, improvements to nvidia-settings, improved Xinerama workstation OpenGL performance, Quad SLI support, improved X driver error recovery, OpenGL 2.1 support, and more. While the 1.0-9XXX series had not delivered a new driver installer or any other groundbreaking changes, the changes that had ended up making it were very nice to see especially with the GLX_EXT_texture_from_pixmap support.

Released a week after the 1.0-9626 driver launch was a security release in the Rel80 series with version 1.0-8776. This driver was brought on by a security hole found within the drivers. See here for more information. The last official driver we will be looking at is the NVIDIA 1.0-9629 release, which mainly publicized the highlights found within the 1.0-9625/1.0-9626 display drivers. This display driver was released on November 7.

Coming out a day after the 1.0-9629 release was the 1.0-9742 beta display driver over at nZone. This driver had offered initial Linux support for the GeForce 8800GTX and 8800GTS. No other changes were officially mentioned. Two legacy drivers had also come out this year -- 1.0-7182 and 1.0-7184. These drivers are solely targeted at the older NVIDIA GPUs, which are no longer supported by the mainstream driver branch. These legacy drivers had added support for X.Org 7.1 and various other fixes.

Note: On the eve of publishing this article, NVIDIA had introduced the 1.0-9631 display driver. This driver is coming out less than a month after the 1.0-9629 release and has three highlights -- fixing an OpenGL crash on some GeForce 3/4 GPUs, fixing an X server crash on some full-screen OpenGL applications, and improved interaction with newer kernels. We were not made aware of this release early by NVIDIA, so we did not have the time to test it for this article. However, this is another bug-fix release so it should have very little effect on the frame-rate performance from the 1.0-9629 or 1.0-9742 display drivers. More information on the 1.0-9631 display driver can be found here.

While feature-wise NVIDIA has made a great deal of improvements with their display drivers this year, one of the areas that can be debated is the quality of their Linux SLI Frame Rendering support. Coincidently the 1.0-8XXX Linux driver series was introduced exactly one year ago to the day. The 1.0-8174 display driver had featured initial support for NVIDIA SLI (Scalable Link Interface) as well as a great deal of other changes. Last year at this time we were here expressing concern over the quality of SLI under Linux. This MultiGPU technology was simply not able to offer any major benefits over a single GPU. With Microsoft Windows the benefits of SLI are apparent for gamers and even workstation users, but with Linux it is simply nonexistent. Even a year later with over a half dozen driver releases, the performance benefits continue to be inferior. With the drivers this year they had introduced Quad SLI support and other fixes, but still we see Linux SLI as unacceptable. Earlier this year we had published an article entitled HP: The SLI Godfather?, in which we had shared that the SLI support under GNU/Linux and Solaris may have solely been because of Hewlett-Packard and Sun Microsystems. As a follow up to that article, we have yet to see any evidence contrary to our investigation and to this date SLI for FreeBSD does not exist.

Latest Articles & Reviews
  1. Intel Broadwell: GCC 4.9 vs. LLVM Clang 3.5 Compiler Benchmarks
  2. Ubuntu vs. Fedora Linux On Lenovo's X1 Carbon With Core i7 Broadwell
  3. Ubuntu 15.04 Is The Easy Path To Better Performance On Intel Broadwell
  4. NVIDIA's Latest Maxwell Line-Up Against AMD With Catalyst On Linux
  5. Preliminary Tests Of Intel Sandy Bridge & Ivy Bridge vs. Broadwell
  6. AMD FX-8320E Performance On Linux
Latest Linux News
  1. I Switched (Back) Over To Fedora As My Main OS & It's Going Great!
  2. Windows 10, PS4, C4 & Systemd News Kicked Off 2015
  3. Calamares 1.0 Distribution-Independent Installer Framework Released
  4. Librem 15 Linux Laptop Set To Close At Around $400k USD
  5. Virtual GEM To Increase Mesa's Software Rasterizer Performance
  6. Open Lunchbox: Yet Another Open-Source Laptop Attempt
  7. Wayland/Weston 1.7 Release Candidate
  8. Bugzilla 5.0 Is Coming Soon With New Features
  9. Linux Benchmarking... Even Faster & A Very Interesting February
  10. Does VirtualBox VM Have Much A Future Left?
Most Viewed News This Week
  1. PlayStation 4 System Compiler Support Landing In LLVM
  2. Does VirtualBox VM Have Much A Future Left?
  3. Linux "GHOST" Vulnerability Hits Glibc Systems
  4. LibreOffice 4.4 Is Coming Soon With New Features
  5. LLVM Adds Options To Do Fuzz Testing
  6. Broadwell Linux Ultrabook Running MUCH Cooler Than Haswell
  7. LibreOffice 4.4 Released With Better OOXML Support, UI Improvements
  8. Inkscape 0.91 Goes Through C++ Code Conversion, New Cairo Rendering, OpenMP Filters