Our training courses give software developers a solid foundation in the latest technologies in GPU computing and provide valuable orientation in current hardware and software. Theory lectures and practical hands-on sessions are employed. Our goal is to give an overview of the basics and benefits of GPU computing technology. In addition, problems and algorithms from the developers’ area of interest may be incorporated into the existing code samples.

Four Days GPU Training Course on CUDA and OpenCL

Nov 30 – Dec 3, 2015
SURFSara Science Park, Amsterdam

Monday (Nov 30) and Tuesday (Dec 1): CUDA

• Introduction to GPU computing
  • Motivation for many-core (GPU) architectures
  • Overview of GPU architectures
• CUDA framework
  • CUDA hardware model
  • CUDA concepts
• CUDA programming
  • Thread management and memory model
  • Optimizations
• Hands-on sessions included.

Wednesday (Dec 2): OpenCL

• OpenCL framework
  • OpenCL programming language
  • OpenCL concepts
• OpenCL programming
  • Threads, memory model and CUDA analogies
  • Optimizations
• Hands-on sessions included.

Thursday (Dec 3): Libraries and Advanced Topics

• GPU libraries
  • Dense matrix libraries
  • Sparse matrix libraries
  • STL-like libraries
• Advanced Optimization
  • Profiling and debugging
  • Multiple / heterogeneous devices
  • Optimizations via texture memory
• Hands-on sessions included.

Goal:

• The developer will have built a foundation for writing basic GPU based applications
• The programmer will be able to further develop their skills by learning GPU example code distributed by vendors
• The programmer will become confident to start developing GPU computing applications

Deliverables: 

• Lecture slides
• Electronic copy of hands-on material
• Certification of completion
• 90 days post training support (conditions apply)
• Lunch and coffee for all four days

Recommendations in C/C++:

• Functions, loops, conditions
• Arrays
• Pointers
• Memory (de)allocation

Pricing and Location:

The total cost for the full course is 2600 EURO (plus VAT if applicable) per attendee. Please contact us, if you would like to participate on specific dates only.

The course will take place at the SURFsara Science Park in Amsterdam.

SURF Science Park Building
Science Park 140
1098 XG Amsterdam
The Netherlands

Hands-on Hardware

All hands-on sessions will be performed on a state-of-the-art cluster equipped with GPU devices.

Registration:

To register for the course, please send a request to training AT paralution.com. The registration deadline is Nov 20, 2015. Space is limited, please register early to guarantee your spot.

You can cancel your registration until Nov 20, 2015. PARALUTION Labs reserves the right to cancel or re-schedule the course up to Nov 20,2015. In both cases, you will have your registration fee fully refunded.

http://cdnc.itec.kit.edu/53_423.php

Contact us for details!