Skip to content. | Skip to navigation

CREATE-MIA!

Personal tools

This is SunRain Plone Theme
You are here: Home / Training / 2015 / GPU Computing with MATLAB Seminars at McGill

GPU Computing with MATLAB Seminars at McGill

GPU Computing with MATLAB Seminars at McGill

What
  • CREATE-MIA Event
  • Technical skills
  • Professional skills
When Jan 30, 2015
from 09:00 AM to 03:00 PM
Where Session 1: Trottier Building, Rm. 2100; Session 2: Trottier Building Room 3120
Attendees All CREATE-MIA Trainees
Attendees from CIM and REPARTI
Add event to calendar vCal
iCal

***

Attendance at this workshop is by registration only.  A registration link will be provided to CREATE-MIA trainees and faculty by email.

***

CREATE-MIA trainees will soon have access to two NVIDIA Tesla K40 GPUs.  This workshop, presented by MathWorks, will  be an introduction to GPU computing using MATLAB with an emphasis on GPU accelerated applications often used in scientific research in computer vision, engineering, and medical image analysis.

Sign-in and Badge Pick-up

9:00am - 9:30am, Trottier Building 2100

Session 1: Optimizing and Accelerating your MATLAB Code

9:30am - 12:30pm, Trottier Building 2100

In this session, we will demonstrate simple ways to improve and optimize your code that can boost execution speed.  We will also address common pitfalls in writing MATLAB code, explore the use of the MATLAB Profiler to find bottlenecks, and introduce programming constructs to solve computationally and data-intensive problems onmulticore computers, clusters and GPUs.

Specifically, we will show:

  • Leveraging the power of vector and matrix operations in MATLAB
  • Identifying and addressing bottlenecks in your code
  • Utilizing additional processing power available in multicore machines, clusters, and grids

Session 2: Developing GPU-accelerated applications with MATLAB

1:00pm - 3:00pm, Trottier Building 3120

Learn how you can use MATLAB to develop GPU-accelerated applications without having to learn the intricacies of GPU architectures or low-level GPU computing libraries.  Following a brief introduction to MATLAB you will work through exercises that show:

  • Using GPU-enabled MATLAB functions to accelerate large matrix operations
  • Minimizing overhead associated with data transfer to the GPU
  • Integrating CUDA kernels in MATLAB

 

This event is being presented in collaboration with  CIM (McGill's Centre for Intelligent Machines) and REPARTI (Regroupement stratégique pour l’étude des environnements partagés intelligents répartis). A huge thank-you to  MathWorks, and the McGill School of Computer Science (SOCS) for making this workshop possible.

« November 2017 »
November
SuMoTuWeThFrSa
1234
567891011
12131415161718
19202122232425
2627282930
Funded by NSERC

Funding provided by NSERC