http://software.intel.com/en-us/articles/3d-vector-normalization-using-256-bit-intel-advanced-vector-extensions-intel-avx en-us http://software.intel.com/en-us/articles/3d-vector-normalization-using-256-bit-intel-advanced-vector-extensions-intel-avx/#comment-49622 Thu, 30 Sep 2010 18:51:59 -0700 http://software.intel.com/en-us/articles/3d-vector-normalization-using-256-bit-intel-advanced-vector-extensions-intel-avx/#comment-49622 ...Newton-Rhapson? The results from _mm_rsqrt_ps() are to a limited amount of accuracy and many applications will refine these results using Newton-Rhapson technique. Yes, using Newton-Rhapson would have shown better scaling over serial. However, an objective was to find a small computation where the aos to soa on-the-fly technique would break-even or be better than serial. Clearly, workloads that do more work than this will further amortize the transpose cost and provide better speedup. ]]> http://software.intel.com/en-us/articles/3d-vector-normalization-using-256-bit-intel-advanced-vector-extensions-intel-avx/#comment-49725 Mon, 04 Oct 2010 12:10:46 -0700 http://software.intel.com/en-us/articles/3d-vector-normalization-using-256-bit-intel-advanced-vector-extensions-intel-avx/#comment-49725 http://software.intel.com/en-us/articles/3d-vector-normalization-using-256-bit-intel-advanced-vector-extensions-intel-avx/#comment-54524 Thu, 06 Jan 2011 06:58:46 -0800 http://software.intel.com/en-us/articles/3d-vector-normalization-using-256-bit-intel-advanced-vector-extensions-intel-avx/#comment-54524 http://software.intel.com/en-us/articles/3d-vector-normalization-using-256-bit-intel-advanced-vector-extensions-intel-avx/#comment-55244 Tue, 18 Jan 2011 17:11:13 -0800 http://software.intel.com/en-us/articles/3d-vector-normalization-using-256-bit-intel-advanced-vector-extensions-intel-avx/#comment-55244 http://software.intel.com/en-us/articles/3d-vector-normalization-using-256-bit-intel-advanced-vector-extensions-intel-avx/#comment-56146 Thu, 10 Feb 2011 15:51:22 -0800 http://software.intel.com/en-us/articles/3d-vector-normalization-using-256-bit-intel-advanced-vector-extensions-intel-avx/#comment-56146 http://software.intel.com/en-us/articles/3d-vector-normalization-using-256-bit-intel-advanced-vector-extensions-intel-avx/#comment-56186 Fri, 11 Feb 2011 08:42:20 -0800 http://software.intel.com/en-us/articles/3d-vector-normalization-using-256-bit-intel-advanced-vector-extensions-intel-avx/#comment-56186 http://software.intel.com/en-us/articles/3d-vector-normalization-using-256-bit-intel-advanced-vector-extensions-intel-avx/#comment-58052 Wed, 30 Mar 2011 23:38:00 -0700 http://software.intel.com/en-us/articles/3d-vector-normalization-using-256-bit-intel-advanced-vector-extensions-intel-avx/#comment-58052 http://software.intel.com/en-us/articles/3d-vector-normalization-using-256-bit-intel-advanced-vector-extensions-intel-avx/#comment-58635 Wed, 13 Apr 2011 10:59:46 -0700 http://software.intel.com/en-us/articles/3d-vector-normalization-using-256-bit-intel-advanced-vector-extensions-intel-avx/#comment-58635 http://software.intel.com/en-us/articles/3d-vector-normalization-using-256-bit-intel-advanced-vector-extensions-intel-avx/#comment-71579 Thu, 15 Mar 2012 07:45:30 -0700 http://software.intel.com/en-us/articles/3d-vector-normalization-using-256-bit-intel-advanced-vector-extensions-intel-avx/#comment-71579