Profile cover photo
Profile photo
Exploiting dataflow parallelism in Teradevice Computing
Exploiting dataflow parallelism in Teradevice Computing

TERAFLUX's posts

Post has attachment
HPCWire: Prof. Guang Gao illustrates advantages of the codelet model

HPCWire: Prof. Guang Gao illustrates advantages of the codelet model. This also investigated together with prof. Gao in the TERAFLUX project.

Post has attachment
Write your program and Run in it in Dataflow mode NOW

This software aims at demonstrating that we can easily provide a very small and powerful runtime for running programs that are coded in whatever programming model, but that could be executed in a DATAFLOW style. The Dataflow Run Time (DRT) provides the runtime support for that
The first benefit of this software is to allow a rapid development of such programs in the context of the TERAFLUX project.

The runtime API has been designed in such way to allow for a future development of a good compiler that targets such interface on one side, and to allow for a good architectural support of such API too (ideally each function could map to an Thread-Level-Parallelism Instruction Set Extension.

Post has attachment
Dataflow Workshop DFM 2013 along with PACT 2013

TERAFLUX contributed to continuing the tradition of the Dataflow Workshp.
This year there has been a reord attendence of about 50 people.
See this website for updates on the proceedings:

TERAFLUX presented at Georgia Tech (Atlanta, GA, USA)

Prof. Roberto Giorgi, scientific coordinator of the TERAFLUX project
presented on 24th August 2013 at Georgia Tech the TERAFLUX project. The event has been kindly organized by prof. Sudhakar Yalamanchili.

TERAFLUX extended to March 31st 2014

The TERAFLUX project had been officially extended to run until March 31st 2014, for a greater impact of the Research results! Prof. Roberto Giorgi (project coordinator).

Post has attachment

Post has attachment
Bridging the Exascale Programming Gap with Dataflow based Technologies

Panel in MULTIPROG, HiPEAC - 21st January 2012 16:00 - Berlin, Germany

The MULTIPROG workshop is hosting a panel organized
by the TERAFLUX coordinator (Prof. Roberto Giorgi)

Dataflow is the most efficient Computational Paradigm. Jack Dennis, the father of Dataflow has defined it as "A Scheme of Computation in which an activity is initiated by presence of the data it needs to perform its function". New ways of exploiting this principle are nowadays available.

In terms of programming models we see at least: OmpSs, OpenStream, DDM-TFLUX, DataFlow-Scala, SWARM, just to name a few. Start-ups like CAPS, MAXELER, are also making available powerful dataflow based technologies. How effective are those technologies? Is there a clear winner? How complementary are those approaches in comparison to other best solutions available in the market?

The invited experts moderated by prof. Roberto Giorgi tried to answer the above and other questions in front of a plubic of about 50 people during the MULTIPROG workshop at HiPEAC conference in Berlin, Germany. The panel of experts included: Prof. Eduard Ayguade - BSC/UPC, Prof. Ian Watson - The University of Manchester, Prof. Albert Cohen - INRIA, Joshua Sutterlein - University of Delaware, Oliver Pell - MAXELER, Lawrance Rauchwerger - TAMU.

Seminar: Control-Flow versus Data-Flow Supercomputers

Veljko Milutinovic, University of Belgrade, presented Seminar: Control-Flow versus Data-Flow SupercomputersVeljko - Siena, Italy, November 12th, 2012
ABSTRACT: This presentation starts with a comparison of various supercomputer types as far as the following issues: (a) Speed, (b) Power, (c) Size, (d) Programming effort, (e) Debugging effort, and (f) Compilation time. It continues with details of the Maxeler approach to data-flow supercomputing, using a number of examples. It concludes with a projection of future trends. If finishes with an elaboration of a PHD research methodology inspired by the scientific success of Maxeler (a spin-off of Stanford and Imperial College London). Data-flow supercomputers compile application code down to the gate level, which helps obtain a number of advantages over Control-flow supercomputers of the same purchasing price. Speedups, for various applications in physics/chemistry/biology, are about 20 times or more, and up to about 200 times for specific business applications, as published by JP-Morgan (a 20% owner of Maxeler). Monthly electricity bills are down for the factor of about 20, which is an important issue, since the two-year electricity bills may overpass the initial investment in the case of Control-flow supercomputers. The size reductions go down also for the factor of about 20. Speedup related data are shown for selected applications in physics, geo-physics, banking, and econometry. A group of PhD-student researchers in Belgrade now develops code for a number of applications not covered so far. They all follow the same methodological path, the details of which will be elaborated in this talk.

Post has attachment
Roberto Giorgi, MATEO'12: Multicore Architectures and Their Effective Operation 2012

Roberto Giorgi presented TERAFLUX project at the Multicore Architectures and Their Effective Operation 2012 -Barcelona, Spain, June 28th - 29th, 2012
Here are the slides of the presentation: TERAFLUX, Effective Operation of Dataflow Parallelism in Teradevices

HPC wire: an article on Prof. Gao and TERAFLUX project
Wait while more posts are being loaded