OpenACC training course

TU Chemnitz | semesterübergreifend OpenACC training course

This training course walks through OpenACC programming technology, from basics to advanced practices used in real applications. As a case study, the course presents a practical session on using OpenACC for solving 2D Poisson problem with FFT. Hands-ons: All discussed topics will be accompanied with practical sessions, using PGI OpenACC compiler with NVIDIA backend. Exercises will be conducted either on the provided remote GPU server or on the customer’s local system. All corresponding presentations will be available to attendees in printed handouts.

CUDA and OpenCL are two programming languages that allow to develop GPU applications with great level of detail, but require advanced knowledge of GPU architecture and extensive code re-writing. Many cross-platform or prototype applications do not actually require deep GPU porting, if they perform data processing using standard algorithms, e.g. reduction. In such cases, fair performance numbers could be achieved with simplified directive-based language extensions. OpenACC is de facto an industry standard for C/Fortran directive extensions for porting code to GPUs.

Applied Parallel Computing LLC is delivering GPU training courses since 2009. Several dozens of courses have been organized all over Europe, both for commercial and academic customers. We work in close partnership with NVIDIA, CUDA Centers of Excellence and Tesla Preferred Partners. In addition to trainings, our company provides GPU porting/optimization services and CUDA certification.

Zugang zum Kurs gesperrt. Bitte melden Sie sich an. Login
Informationen zum Zugang
Sie haben zu wenig Berechtigungen, um diesen Kurs zu starten.