Main Profile

At A Glance

Evaluating MapReduce for Multi-core and Multiprocessor...

Google Video tech TalksFebruary 27, 2007ABSTRACTThis paper evaluates the suitability of the MapReduce model for multi-core and multi-processor systems. MapReduce was created by Google for application development on data-centers with thousands of servers. It allows programmers to write functional-style code that is automatically parallelized and scheduled in a distributed system.We describe Phoenix, an implementation of MapReduce for shared-memory systems that includes a programming API and an efficient runtime system. The Phoenix runtime automatically manages thread creation, dynamic task scheduling, data partitioning, and fault tolerance across processor nodes. We study Phoenix with...
Length: 44:47

Contact

Questions about Evaluating MapReduce for Multi-core and Multiprocessor...

Want more info about Evaluating MapReduce for Multi-core and Multiprocessor...? Get free advice from education experts and Noodle community members.

  • Answer

Ask a New Question