/samplematrixcode

Sample matrix multiply code to show affect of blocking and data alignment

Primary LanguageC

samplematrixcode

Sample matrix multiply code to show affect of blocking and data alignment The code mm.c accompanies two papers at software.intel.com that discuss memory layout and performance. A simple matrix multiply is reordered and blocked to show performance improvement An exercise is included to show the impact on performance when matrices are not aligned on cacheline boundaries.