We would like to bring your attention to the https://software.intel.com/en-us/vtune-amplifier-cookbook-false-sharing recipe from the Performance Analysis Cookbook. It demonstrates on a real example how VTune Amplifier can be used to detect and fix memory issues related to contended access - in particular false sharing.
Finding and especially resolving micro-architecture performance issues could be challenging. We are hoping that sharing such examples will help you with this.