Intel SNA Acceleration Performance On Ironlake

Written by Michael Larabel in Intel on 15 December 2011 at 11:07 AM EST. 5 Comments
INTEL
Wondering how Intel's SNA acceleration architecture is performing for Ironlake hardware? Here's some benchmarks.

Following the recent release of the xf86-video-intel 2.17 X.Org driver I delivered some new SNA benchmarks on Sandy Bridge hardware. There's been a continual flow of changes for Sandy Bridge New Acceleration since the crazy fast acceleration architecture was introduced back in June.

While Intel's focus is now on Sandy Bridge and next-generation Ivy Bridge hardware, SNA acceleration does have back-ends for previous generations of Intel integrated graphics. The last time I delivered SNA benchmarks on non-Sandy-Bridge hardware was back in September, but last weekend I ran some updated tests with the latest Git code for xf86-video-intel, Mesa, and the Linux kernel.

As the 2.17 tests on Sandy Bridge showed, some of the SNA performance gains have been erased as the architecture has matured and been fixed up in various areas. However, unfortunately, the state of SNA right now for previous-generation Ironlake (Clarkdale/Arrandale CPUs) isn't too good.

My experience with the Git code as of last weekend was quite poor. When running fullscreen OpenGL benchmarks the system would end up locking-up when using the Intel DDX driver with SNA enabled. When it came to the 2D benchmarks, there were a few stability issues too, but at least tests could finish successfully.

The testing was being done from an ASRock NetTop with an Intel Core i3 330M processor. The mixed results from this SNA on Ironlake testing can be analyzed on OpenBenchmarking.org. Embedded below are just a couple of the 2D test results comparing the same xf86-video-intel driver with and without SNA support.
Intel Ironlake SNA 2D

Intel Ironlake SNA 2D

Intel Ironlake SNA 2D

Intel Ironlake SNA 2D

Intel Ironlake SNA 2D

Utilizing Sandy Bridge New Acceleration on any generation of Intel graphics currently isn't the default and requires building the DDX with the --enable-sna compile-time switch.

The support for SNA isn't perfect, but Chris Wilson is actively working on it, as can be seen from the Git log. Even since the tests this past weekend there's been many SNA-related commits to master by Chris. Even over the weekend he implemented a VMA cache for SNA thanks to the Phoronix testing.

Find the rest of the results on OpenBenchmarking.org. You can also compare your system's 2D performance to these Intel Ironlake numbers by simply running phoronix-test-suite benchmark 1112104-AR-INTELIRON40 from your system, regardless of whether it's Linux, BSD, or Solaris for these X.Org benchmarks.
Related News
About The Author
Michael Larabel

Michael Larabel is the principal author of Phoronix.com and founded the site in 2004 with a focus on enriching the Linux hardware experience. Michael has written more than 20,000 articles covering the state of Linux hardware support, Linux performance, graphics drivers, and other topics. Michael is also the lead developer of the Phoronix Test Suite, Phoromatic, and OpenBenchmarking.org automated benchmarking software. He can be followed via Twitter, LinkedIn, or contacted via MichaelLarabel.com.

Popular News This Week