This shows you the differences between two versions of the page.

Link to this comparison view

news:timo_bingmann_visiting [2018-10-02 13:02] (current)
Florian Kurpicz created
Line 1: Line 1:
 +====== Timo Bingmann visiting ======
 +Timo Bingmann from KIT will be visiting from October 10 to October 11 and give a talk on "//​Thrill:​ High-Performance Algorithmic Distributed Batch Data
 +Processing with C++//" in OH14 R202 on October 10 at 14:00.
 +**Abstract** - We present on-going work on a new distributed Big Data processing
 +framework called Thrill. It is a C++ framework consisting of a set of
 +basic scalable algorithmic primitives like mapping, reducing, sorting,
 +merging, joining, and additional MPI-like collectives. This set of
 +primitives goes beyond traditional Map/Reduce and can be combined into
 +larger more complex algorithms, such as WordCount, PageRank, k-means
 +clustering, and suffix sorting. These complex algorithms can then be run
 +on very large inputs using a distributed computing cluster. Among the
 +main design goals of Thrill is to lose very little performance when
 +composing primitives such that small data types are well supported.
 +Thrill thus raises the questions of a) how to design algorithms using
 +the scalable primitives, b) whether additional primitives should be
 +added, and c) if one can improve the existing ones using new ideas to
 +reduce communication volume and how they can be scaled to very large
 +cluster sizes.
Last modified: 2018-10-02 13:02 by Florian Kurpicz