Skip To Content
CUDA Workshop is a Course

CUDA Workshop

Aug 30 - Aug 30, 2019
4.0 credits

Enroll

Full course description

Term: Fall 2019
Date: August 30th, 2019
Time: 8:00am - 12:00pm
Location: Torgerson 1100
Instructor: Robert Settlage
Presented By: Advanced Research Computing

Description:
Writing CUDA kernels that take full advantage of the massively parallel hardware can be a challenge. This class will cover common pitfalls and optimization techniques for Nvidia’s programming language.

Topics covered include:

– Memory bound vs compute bound kernels

– Selecting the best threads/blocks combination to achieve full occupancy in a device

– Shared memory access patterns

– Instruction level parallelism

– Using templates in CUDA

Sign up for this course today!

Enroll