ICTS is organizing a series of weekly meetings with focus on High Performance Computing (HPC) and parallel programming. The idea is to combine basic topics and their possible applications in the main scientific areas of interest for the ICTP Community.
Alternating, HPC specialists from the ICT Section will present one or more topics within a self-contained one-hour lecture. Therefore, people interested in one argument, can participate in the specific lecture without the necessity to follow the whole program.
The sessions will be organized every Tuesday from 11:00 to 12:15 starting on October 9, 2012 until the end of this year. Exceptions will be:
- Wednesday November 14, 2012 10:00AM - 11:15 at Leonardo Da Vinci Building, Stasi Seminar Room
- Wednesday November 21, 2012 10:00AM - 11:15 at Leonardo Da Vinci Building, Stasi Seminar Room
Below you can find the complete scheduled program. The page will further be upgraded as needed. Lecture slides will also be by time to time uploaded and made available for download. Schedule and lecture rooms might change too.
This web page along with both the science-ts@lists.ictp.it and the allusers@lists.ictp.it mailing-lists are the networks where news will get around among. Stay tuned!
Speakers |
Date |
Location |
Subject |
Documentation |
Ivan Girotto |
October 9, 2012 11AM - 12:15 |
Leonardo Da Vinci Building, Euler Lecture Hall |
Thinking in Parallel. Brief description: domain decomposition and task parallelism - basic concept of parallelism applied to a common problems (FFT, GEMM and more) |
Thinking in Parallel |
Ivan Girotto |
October 16, 2012 11AM - 12:15 |
Leonardo Da Vinci Building, Euler Lecture Hall |
How to evaluate the efficiency of the parallelism. Brief description: Speedup, efficiency and scalability.
Understanding how to perform efficient simulations on parallel
architecture. |
How to Evaluate the Efficiency of The Parallelism |
Axel Kohlmeyer |
October 23, 2012 11AM - 12:15 |
Leonardo Da Vinci Building, Euler Lecture Hall |
Introduction to HPC. Brief description: why do we are and what is it about. Overview of the main concepts who are behind the fancy label "HPC". |
Introduction to HPC, why do we are and what is it about. |
David Grellsheid | October 30, 2012
11AM - 12:15 |
Leonardo Da Vinci Building, Euler Lecture Hall | HPC in particle
physics. Brief description: Computational tasks in High Energy Physics. Overview of current strategies, and the features and problems of the LHC Computing Grid. |
Computing in Particle Physics. |
Axel Kohlmeyer |
November 6, 2012 11AM - 12:15 |
Leonardo Da Vinci Building, Euler Lecture Hall |
HPC for science. Brief description: Axel will go through the most interesting experiences of his long career showing how HPC have had a significant impact for making real science. |
|
Axel Kohlmeyer |
November 14, 2012 |
Leonardo Da Vinci Building, Stasi Seminar Room |
Fighting errors when
either compiling/installing or running
software. |
|
Ivan Girotto | November 21, 2012 10:00AM - 11:15 |
Leonardo Da Vinci Building, Stasi Seminar Room | Parallel Computing Architecture: how to exploit massive
parallel system such as BG/Q. Brief description: application tuning on massive parallel systems, the IBM Blue Gene/Q. Examples: Quantum ESPRESSO & RegCM. |
Tuning of Scientific Applications on the IBM BG/Q Architecture |
Axel Kohlmeyer | November 27, 2012 11AM - 12:15 |
Leonardo Da Vinci Building, Euler Lecture Hall | Language relations. How to access C/C++ from Fortran
and the other way around? Brief description: How to recognize "mangled" function names when profiling or tracing (gprof, perf, valgrind/memcheck). |
Language relations. How to access C/C++
from Fortran and the other way around? |
Axel Kohlmeyer | December 4, 2012 11AM - 12:15 |
Leonardo Da Vinci Building, Euler Lecture Hall | Lennard-Jones
code for Molecular Dynamics: from serial to
multithread. Brief description: a practical example of using multi-threading programming paradigm for parallel development. |
|
Axel Kohlmeyer | December 11, 2012 11AM - 12:15 |
Leonardo Da Vinci Building, Euler Lecture Hall | Lennard-Jones
code for Molecular Dynamics: from multi-threading to distributed
memory. Brief description: a practical example of using mixed messsage passing and multi-threading programming paradigms for parallel development. |
|
Luca Tornatore (SISSA) | December 18, 2012 11AM - 12:15 |
Leonardo Da Vinci Building, Euler Lecture Hall |
HPC in Astrophysics Brief description: how to compute
a universe in a box. A journey through massively parallel codes in
cosmology, from MPI-multithreading mixture to computational
strategies and algorithms. |
|