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

A 13 Line Patch That Boosts Intel Sandy Bridge Performance

Michael Larabel

Published on 4 March 2011
Written by Michael Larabel
Page 1 of 4 - 25 Comments

After some initial Linux troubles, last month we finally got Intel Sandy Bridge graphics working under Linux. The latest Intel CPUs (such as the Core i5 2500K) with integrated graphics are blazingly fast, and the classic Intel Mesa driver was fast compared to other open-source Mesa / Gallium3D drivers, but it still was a ways behind the low-end discrete graphics cards with the proprietary AMD / NVIDIA drivers for Linux. It was also shown that the Intel Linux Mesa driver is much slower than the Intel Windows driver for Sandy Bridge, as we had also found was the case for previous generations of Intel graphics. Committed to the Mesa mainline Git repository this week though was a very important Sandy Bridge change. While the commit only touched 13 lines of code (11 lines of new code, 2 lines of changed code), it has dramatically improved the Sandy Bridge Linux performance as our results show in this article.

It is this Git commit from the 1st of March by Intel's Zou Nan hai that we are talking about. For Sandy Bridge / "Gen 6" Intel hardware, it bumps the VS thread count to 60. It is a relatively straightforward and simple patch, but how it affects the OpenGL performance is dramatic.

In this article are the original Phoronix test results for the Intel Core i5 2500K "Sandy Bridge" from early February using the Intel Bearup Lake motherboard (the show-stopping problem with the ASUS motherboard remains unresolved) plus the already-published Windows 7 Professional x64 Service Pack 1 results. What is new are the updated Intel numbers. On a clean install of Ubuntu 10.10 to the same Sandy Bridge system, the latest Linux 2.6.38 kernel was loaded (2010-03-02; Linux 2.6.38-rc7) plus the latest xf86-video-intel DDX from Git on 2010-03-02 (xf86-video-intel 2.14.901), the latest Git libdrm, and then Mesa from Git master on 2010-03-02 marked as Mesa 7.11-devel.

Besides the Intel Core i5 2500K processor and Intel Bearup Lake motherboard was 2GB of DDR3 system memory and a 60GB OCZ Vertex 2 SSD. Via the Phoronix Test Suite 3.0 and OpenBenchmarking.org software, the test results from Nexuiz, OpenArena, World of Padman, Warsow, Urban Terror, Tremulous, VDrift, and Lightsmark were used.

<< Previous Page
1
Latest Linux Hardware Reviews
  1. NVIDIA GeForce GTX 970 Offers Great Linux Performance
  2. CompuLab Intense-PC2: An Excellent, Fanless, Mini PC Powered By Intel's i7 Haswell
  3. From The Atom 330 To Haswell ULT: Intel Linux Performance Benchmarks
  4. AMD Radeon R9 285 Tonga Performance On Linux
Latest Linux Articles
  1. 6-Way Ubuntu 14.10 Linux Desktop Benchmarks
  2. Ubuntu 14.10 XMir System Compositor Benchmarks
  3. Btrfs RAID HDD Testing On Ubuntu Linux 14.10
  4. Ubuntu 14.10 Linux 32-bit vs. 64-bit Performance
Latest Linux News
  1. Mono Brings C# To The Unreal Engine 4
  2. Coreboot Now Has Support For Intel Broadwell Hardware
  3. Enlightenment's EFL 1.12 Alpha Has Evas GL-DRM Engine, OpenGL ES 1.1 Support
  4. GTK+ Lands Experimental Backend For Mir Display Server
  5. Ubuntu 14.10 Officially Released
  6. Mesa 10.4 Might Re-Enable HyperZ For R600g/RadeonSI
  7. Intel GVT-g GPU Virtualization Moves Closer
  8. GTK+ 3.16 To Bring Several New Features
  9. Debian 8.0 Jessie Has Many Multimedia Improvements
  10. What Linux Benchmarks Would You Like To See Next?
Latest Forum Discussions
  1. Linux hacker compares Solaris kernel code:
  2. HOPE: The Ease Of Python With The Speed Of C++
  3. Advertisements On Phoronix
  4. Updated and Optimized Ubuntu Free Graphics Drivers
  5. Users/Developers Threatening Fork Of Debian GNU/Linux
  6. Ubuntu 16.04 Might Be The Distribution's Last 32-Bit Release
  7. AMD Releases UVD Video Decode Support For R600 GPUs
  8. Proof that strlcpy is un-needed