AMD Bulldozer Review, AVX Performance using SONAR benchmark

Its pretty cool that Cakewalks AVX optimization work was featured in this review of AMD’s bulldozer from Tom’s Hardware. For those unfamiliar with Tom’s Hardware, the site is the holy grail of hardware reviews and benchmarks. Their depth of knowledge and coverage of the state of the art in computer hardware is unparalleled.

A few months ago I was contacted by Chris Angelini, the Editor in chief at Tom’s Hardware. He had come across a white paper that I co authored with Intel, featuring the AVX optimization’s in SONAR X1, and was very interested in knowing more about our experiences with AVX in relation to Bulldozer vs Intel’s SandyBrige.

Since we didn’t have access to the the newer AMD systems with this chipset, I sent Chris a copy of SONAR X1 so that he could run some comparative studies himself. I also sent him a small bench-marking application that we had developed in-house to use as a test harness for the AVX optimized bits of code covered in the white paper. Using this app he was able to objectively test the performance of the AVX instruction set on the AMD and Intel platforms using code from SONAR X1. You can read the results of his tests in the links below.

Working with Chris on this, I was really blown away with his depth of knowledge and the exhaustive level of testing they do with this stuff at Tom’s Hardware. Its great seeing our software being used to validate performance in a mainstream CPU benchmark review like this!

Links:

Review referencing Cakewalk AVX tests
full 24 page review
Intel/Cakewalk AVX Whitepaper
SONAR X1

2 Replies to “AMD Bulldozer Review, AVX Performance using SONAR benchmark”

  1. It seems one of the AVX optimised routines (Copy Float32toFloat64) has a problem on the Bulldozer. As this test was performed almost a year ago and the reviewer suggests an application patch or Visual Studio service pack might be needed to fix the problem, it would be interesting to get an update.

    Is Sonar smart enough to detect that the AVX specific operation is 77% slower and go back to the conventional non-AVX process so as not to incur such a big hit? Or has this problem been fixed with a patch?

  2. I’ve been listening to the voices on the cakewalk forum and I must say there have been a lot of great story’s of users using AMD Phenom’s 4/6 and 8 core CPU’s
    I myself own two, a 4 core and a 6 core.

    the speed, the power, its incredible.
    Can you tell me why?

    what is it that makes the AMD phenom series solid as a rock???
    this is not jus one user who appreciates AMD.
    this is months of collected data on users complaints/positive feedback.
    they just make sense.

Leave a Reply to Bill Ruys Cancel reply

Your email address will not be published. Required fields are marked *