Portable GPU accelerated computing with C++ and HIP - Private Group Online class
This is a three day hands-on course into accelerating C++ applications using HIP. The course is taught live, remotely online, and scheduled at your organization’s convenience. Attendees get access to Fluid Numerics’ AMD & Nvidia GPU systems and remote office hours for a full week to work through hands-on exercises.
In this course, we cover the following topics
GPU Computing Fundamentals
Building HIP applications with CMake
Host and device memory management in HIP applications
Asynchronous operations using streams
Debugging HIP applications using roc-gdb
Profiling applications using rocprof, omniperf, and omnitrace
Basic kernel optimization strategies
Porting CUDA applications to HIP
Class size is limited to 12 students.
This is a three day hands-on course into accelerating C++ applications using HIP. The course is taught live, remotely online, and scheduled at your organization’s convenience. Attendees get access to Fluid Numerics’ AMD & Nvidia GPU systems and remote office hours for a full week to work through hands-on exercises.
In this course, we cover the following topics
GPU Computing Fundamentals
Building HIP applications with CMake
Host and device memory management in HIP applications
Asynchronous operations using streams
Debugging HIP applications using roc-gdb
Profiling applications using rocprof, omniperf, and omnitrace
Basic kernel optimization strategies
Porting CUDA applications to HIP
Class size is limited to 12 students.
This is a three day hands-on course into accelerating C++ applications using HIP. The course is taught live, remotely online, and scheduled at your organization’s convenience. Attendees get access to Fluid Numerics’ AMD & Nvidia GPU systems and remote office hours for a full week to work through hands-on exercises.
In this course, we cover the following topics
GPU Computing Fundamentals
Building HIP applications with CMake
Host and device memory management in HIP applications
Asynchronous operations using streams
Debugging HIP applications using roc-gdb
Profiling applications using rocprof, omniperf, and omnitrace
Basic kernel optimization strategies
Porting CUDA applications to HIP
Class size is limited to 12 students.