NVIDIA GPU Programming Workshop
September 29-30 2015, 9AM-5PM at the Graduate Center
NVIDIA, the CUNY HPC and the CUNY Graduate Center are pleased to offer a two-day GPU programming workshop, September 29-30 2015
LOCATION: The Graduate Center
The City University of New York
365 5th Avenue,
New York, NY 10016
A simple lunch will be provided each day. There is no charge for this workshop.
During the 2 days, attendees will be exposed to 3 major approaches to adding GPU acceleration to their applications, along with hands-on programming examples, and basic optimization techniques.
Tuesday, September 29, 2015
1. Intro to GPU computing
2. Intro to CUDA C/C++
3. CUDA basic optimizations
4. Working with GPU Libraries (CUBLAS, CUSPARSE)
5. Brief intro to CUDA Fortran
Wednesday, September 30, 2015
1. Intro to OpenACC
2. OpenACC basic optimizations
3. Demo of GPU Graphical tools: Profiler, Nsight EE, debugging
4. Analysis-driven optimization
Attendees who wish to participate in the CUDA programming exercises are expected to bring their own laptop with Wifi connectivity, and should have a suitable SSH client installed, such as PuTTY for Windows users.
Familiarity with basic Linux command line operations and usage of a basic text editor such as vi will be beneficial.
Exercises are conducted primarily using the C and C++ programming languages.
At the start of the workshop, attendees will receive instructions on setting up a temporary account on the CUNY HPC Center GPU computing resources that will be used to conduct the workshop hands-on exercises.
Seating is limited and registration is required.
To register please send an email to: HPCWorkshops@csi.cuny.edu
And include the following information:
1. First and Last Name
2. Your CUNY Email Address
3. Academic Status (ie:
(Faculty, Researcher, PostDoc, Graduate Student, Undergrad)
4. CUNY Campus
5. Department / Research Field
Parallel Programming and Optimization with
Intel Xeon Phi Coprocessors
Developer Training Event
at the CUNY Graduate Center
Tuesday, October 20, 2015
The Graduate Center
4th Floor - Science Center
356 Fifth Avenue
New York, NY 10016
Reservations are required as seats are limited.
This one-day training features presentations and hands-on exercises on the available programming models and best optimization practices for the Intel Xeon Phi coprocessors, and on the usage of the Intel software development and diagnostic tools.
- Offload and Native: "Hello World" to complex, using MPI.
- Case Study: All aspects of tuning in the N-body calculation.
- Optimization I: Strip-mining for vectorization, parallel reduction.
- Optimization II: Loop tiling, thread affinity.
- Intel Xeon Phi architecture: purpose, organization, prerequisites for good performance, future technology
- Programming models: native, offload, heterogeneous clustering
- Parallel frameworks: automatic vectorization, OpenMP, MPI
- Optimization methods: general, scalar math, vectorization, multithreading, memory access, communication and special topics
*Please bring your own laptop to the training; below is a list of the necessary specifications:
- Windows (XP or newer), Mac OS X (10.5 or later), or Linux (something from the 21st century)
- Wired (Ethernet) and wireless (Wi-Fi 802.11g or later) network connectivity
- Web Browser (any except Microsoft Edge)
- On Windows: Putty and Pageant (www.putty.org) and WinSCP (www.winscp.net)
- On Mac OS X and Linux: ssh client
- Optional: on all operating systems, the free software NoMachine (www.nomachine.com).
This is only necessary if you are not comfortable programming in Linux in a text terminal over an SSH connection.
To register for the workshop, please click here: Register for Oct Training