Jeff's Notebook: Benefits of Intel® Advanced Vector Extensions (Intel® AVX) for Small Matrices

With the launch of the Second Generation Intel® CoreTM processor (utilizing the new Intel® processor microarchitecture codename Sandy Bridge), Intel released Intel Advanced Vector Extensions (Intel AVX) which extends the capabilities of Intel® Streaming SIMD Extensions (Intel® SSE), especially for floating point data and operations. Intel AVX significantly increases the floating-point performance density with improved power efficiency over previous 128-bit SIMD instruction set extensions.

Recently, I found an interesting whitepaper, "Benefits of Intel AVX for Small Matrices",  written by Pallavi Mehrotra and Richard Hubbard that examines how some of the Intel AVX architecture features such as wider 256-bit registers, new data manipulation and arithmetic primitives can benefit operations on matrices of small sizes. Their whitepaper examines simple operations such as matrix addition and multiplication, as well as more complex ones like computing the determinant of a matrix. Read their whitepaper to get the full details.

Einzelheiten zur Compiler-Optimierung finden Sie in unserem Optimierungshinweis.