Centro de Matemática Aplicada y Computo de Alto Rendimiento
ABACUS-CINVESTAV EDOMEX - México
28th of june to 4th of july, 2015
SPONSORS
MEDIA PARTNERS
SC-Camp Archive
SC-Camp 2010
SC-Camp 2011
Program
|
| Shared and Distributed Memory | ||
| Time | Description | |
| 07:00 - 08:00 | Breakfast | |
| 08:00 - 09:30 | Introduction to Distributed Memory Models | |
| 09:30 - 11:00 | Distributed Computing Through Message Passing MPI | |
| 11:00 - 12:00 | Introduction to Shared Memory Models | |
| 12:00 - 13:00 | OpenMP Essencials | |
| 13:00 - 14:00 | How to Compute Discrete Logarithms over a 4841-bit Ternary Extension Field in Less Than Two Hundred Twenty Core-years | |
| 14:00 - 16:00 | Lunch | |
| 16:00 - 19:00 | Practical Session : MPI+OpenMP | |
| 19:00 - 20:00 | Diner | |
Introduction to Distributed Memory Models
08:00 - 09:30
Speaker: Gilberto Días
Programming models to understand the trands when using multiple computing units.
Distributed Computing Through Message Passing MPI
09:30 - 11:00
Speaker: Gilberto Días
The challanges of distributed computing involve exchanging message. The MPI, message passing interface, is still THE standard for many distributed applications. Throughout this lecture we will visit many aspects of distributed programing usign MPI.
Introduction to Shared Memory Models
11:00 - 12:00
Speaker: Robinson Rivas
Todays' HPC systems always count on multiple cores. With the goal to improve the usage on such systems knowledge of multithread programing is mandatory. This lectures shows the basis of OpenMP : a well spred environment to implement parallel applications.
OpenMP Essencials
12:00 - 13:00
Speaker: Robinson Rivas
Todays' HPC systems always count on multiple cores. With the goal to improve the usage on such systems knowledge of multithread programing is mandatory. This lectures shows the basis of OpenMP : a well spred environment to implement parallel applications.
How to Compute Discrete Logarithms over a 4841-bit Ternary Extension Field in Less Than Two Hundred Twenty Core-years
13:00 - 14:00
Speaker: Francisco Rodriguez
In the past two years, there have been several dramatic improvements in algorithms for computing discrete logarithms in small-characteristic finite fields. In this talk, we examine the effectiveness of these new algorithms for computing discrete logarithms in the field GF(36*509). The intractability of the discrete logarithm problem in this field is necessary for the security of bilinear pairings derived from supersingular curves with embedding degree 6 defined over GF(36*509). These curves were believed to enjoy a security level of 128 bits against attacks by the classical Coppersmith's algorithm. Our analysis shows that these pairings offer security levels of at most 56.9 bits. In this talk we present a work in-progress report of the computation of discrete logs in the field GF(36*509). Our computation is divided in 4 major steps, each of them having a complexity of approximately 256 bits. Taking advantage of the highly parallelizable nature of the algorithm, we have deployed the algorithm execution in a cluster of 270 nodes. After more than two months of computation (equivalent to roughly 50 core-years of computation), we have managed to obtain the first results for the first two major steps of the algorithm. At this rate, we hope to finish the computation in six more months. When finished, we will compute a discrete log over a 4841-bit field, which would break the current world record of 4404 bits established in 2014 by a Swiss team.
Practical Session : MPI+OpenMP
16:00 - 19:00
Speaker: Gilberto Días and Robinson Rivas
Using MPI and OpenMP we will guide you on the implementation of a parallel and distributed application.