Parallel in R
Details
The first 2015 SRUG meetup will be on parallel computing in R. Knowing how to program in parallel is becoming increasingly important as today’s gains in computational power are mostly due to increases in the available cores. Fortunately for, it is easy to tap into this resource in R. This session will give an introduction to both the basics and the more advanced aspects of parallelization. The topics are:
Introduction to parallel programming in R (Max Gordon)
There will be a short introduction to general concepts such as split-apply-combine and the MapReduce model. After this we will explore the parallel package, the foreach package, how to deal with common issues such as debugging, memory handling, caching, etc.
Going big with Spark (Mikael Huss)
When your local laptop is insufficient for your task there are nowadays a wide variety of ways to do your computations on large clusters. While the base theory of parallelization tasks is unchanged, there are new challenges when combining hundreds of computers into one computational unit. In order to overcome this bridge there are popular utilities that aim at providing the end user with a simple interface, the perhaps most recent addition is Spark. We will try to cover the basics: where to find it, how to set it up, and some personal experience from the process.
Big data juggling with dplyr (Love Hansson)
With dplyr, semantic data management has finally made its way into the R world. Like data.table, dplyr makes heavy use of under-the-hood parallellization which has led to serious performance gains compared to base-R data management.
This talk will focus on how dplyr can be used for exploration, fast prototyping and production when working with data flows, large or small.
******************************************
Chance to win conference pass tonight
******************************************
If you come tonight you can win a free pass to Strata+ Hadoop World
It's a gift to SRUG from
SRUG members receive a 20% discount to Strata + Hadoop World 2015 by using this link
and your unique code*: SRUG20
Parallel in R