INTRODUCTION

Data sharing - a widely accepted model of communication between parallel threads

HW support: shared memory multiprocessor

Private cache memories:

• to speed-up memory access time
• to decrease contention

Typical architecture “Dance-Hall”