2014 WORKSHOPS & TUTORIALS

The 11th International Meeting on High Performance Computing for Computational Science (VECPAR 2014) program will include one workshop and two tutorials, Scheduled for June 30 and July 1, 2014.

Workshops

Tutorials




Workshops



W1: The Ninth International Workshop on Automatic Performance Tuning (iWAPT)

Duration and schedule:

Full Day Workshop (Tuesday, July 1 8:30 - 17:00)

Location:

Eugene Hilton

Workshop Chair:

Franz Franchetti, Carnegie Mellon University, USA

Description:

The goal of the Ninth International Workshop on Automatic Performance Tuning (iWAPT2014) is to bring together researchers who are investigating automated techniques for constructing and/or adapting algorithms and software for high-performance on modern complex machine architectures.

We are particularly interested in autotuning and its relationship to the following topic areas, among others:

Workshop's web page

iWAPT 2014 Program

Tuesday, July 1st

08:30 − 10:00Session 1
8:30 – 9:00 Invited Talk: Boyana Norris
The end of coding as we know it
Faced with increasingly parallel, heterogeneous, and diverse architectures, developers are faced with a growing number of tradeoffs they must consider. What data structure is best for a given computation? What happens when that same data must then be accessed in a different manner by another algorithm? How can different modules be coupled most effectively? Which library and specific function is best for a given subcomputation? What parallel programming model should I use? At present, most of these choices are made based on the developers’ expertise and experience and subject to many external constraints, such as upcoming deadlines, available human and computational resources, and many other human factors. As a result, the achieved solution rarely achieves the full potential of the algorithm or underlying architectures. Autotuning approaches offer a means to procrastinate from making these choices and delegate them instead to one or more tools that attempt to generate and fine-tune implementations automatically for a given target platform. In this talk I will present our autotuning approach and discuss its place in the current ontology of autotuning approaches, and discuss future autotuning/code generation opportunities.
9:00 – 9:30 Yu Lin, Franjo Ivancic, Pallavi Joshi, Gogul Balakrishnan, Malay Ganai and Aarti Gupta
Environment-Sensitive Performance Tuning for Distributed Service Orchestration Paper PDF
9:30 – 10:00 Shahzeb Siddiqui and Saber Feki
Historic Learning Approach for Auto-tuning OpenACC Accelerated Scientific Applications Paper PDF
10:30 − 12:00Session 2
10:30 – 11:00 Invited Talk: Keita Teranishi
Auto-Tuning for Unreliable HPC
11:00 – 11:30 Georg Ofenbeck, Tiark Rompf, Alen Stojanov, Martin Odersky and Markus Püschel
Building Program Generators for High-Performance: Spiral on Scala Paper PDF
11:30 – 12:00 Richard Veras and Franz Franchetti
Capturing the Expert: Generating Fast Matrix-Multiply Kernels with Spiral Paper PDF
13:30 − 15:00Session 3
13:30 – 14:00 Elmar Peise and Paolo Bientinesi
A Study on the Influence of Caching: Sequences of Dense Linear Algebra Kernels Paper PDF
14:00 – 14:30 France Boillod-Cerneux, Serge Petiton, Christophe Calvin and Leroy Drummond
Toward Restarting Strategies Tuning for a Krylov Eigenvalue Solver Paper PDF
14:30 – 15:00 Takeshi Fukaya, Toshiyuki Imamura and Yusaku Yamamoto
Performance Analysis of the Householder-type Parallel Tall-Skinny QR Factorizations toward Automatic Algorithm Selection Paper PDF
15:30 − 17:00Session 4
15:30 – 16:00 Takeshi Minami, Motoharu Hibino, Tasuku Hiraishi, Takeshi Iwashita and Hiroshi Nakashima
Automatic Parameter Tuning of Three-Dimensional Tiled FDTD Kernel Paper PDF
16:00 – 16:30 Nicholas Chaimov, Boyana Norris and Allen Malony
Integration and Synthesis for Automated Performance Tuning: the SYNAPT Project Paper PDF
16:30 – 17:00 Alfian Amrizal, Shoichi Hirasawa, Hiroyuki Takizawa and Hiroaki Kobayashi
Automatic Parameter Tuning of Hierarchical Incremental Checkpointing Paper PDF


Tutorials



T1: Programming and Optimizing for the Intel® Xeon Phi™ Coprocessor

Duration and schedule:

Half Day Tutorial (Monday, June 30)

Location:

Eugene Hilton

Organizers:

Martyn J. Corden, Developer Products Division, Intel, Hillsboro, OR, USA

Description:

Come and learn about programming models and optimization for Intel® Xeon Phi™ Coprocessors. Programming for native execution is just like the familiar programming for Intel® Xeon® processors and programming for offload from an Intel® Xeon® host processor using OpenMP 4.0 is a straightforward extension. The tutorial discusses various optimization techniques with a particular emphasis on vectorization. Careful optimization is even more important for getting good performance on Intel® Xeon Phi™ Coprocessors, but uses the same underlying techniques as optimizing for the host processor. Hands-on examples and labs illustrate vectorization, OpenMP threading and random number generation in the coprocessor environment, including performance analysis using Intel® VTune™ Amplifier XE.

A laptop with ssh or PuTTY will be a requirement in order to do any hands on work. VNC will probably be required in order to use VTune, but people without will still be able to compile, run and time codes in native and offload modes.

Tutorial Overview Slides
VTune and Amplifier Intro Slides


T2: Trilinos

Duration and schedule:

Full Day Tutorial (Monday, June 30)

Location:

Eugene Hilton

Organizers

Mark Hoemmen, Sandia National Labs, USA

Description:

Trilinos (trilinos.org) is a software library for solving large-scale mathematical problems arising in science and industry. It can solve enormous problems on very large parallel computers, yet can also run on your laptop. Trilinos includes mathematical objects like sparse matrices and dense vectors; solvers for linear, nonlinear, and optimization problems; preconditioners; tools for discretizing conventional and stochastic partial differential equations; time integrators; automatic differentiation; and many other algorithms and enabling technologies. This tutorial will begin with a broad overview of what Trilinos can do for you. You can then try it for yourself through a guided hands-on exercise.

Trilinos Tutorial Slides PPTX
Kokkos Overview Slides PPTX
Kokkos Tutorial Slides PPTX