High Performance Computing: Programming GPU using CUDA Fortran

PrerequisitesNew

The course aims to give a very introductory overview of GPU Programming and CUDA FORTRAN language.

• All current Cambridge University members (departments and colleges)
• Especially University alumni (PhD) and researchers who already have access to HPCS computing facilities
• Further details regarding eligibility criteria are available

• Basic FORTRAN77/FORTRAN90 knowledge is required
• Basic knowledge of Unix environment
• No required knowledge of GPU computer is needed

Number of sessions: 2

• CUDA FORTRAN history
• Crash introduction to GPU Computing and CUDA C
  • memory model
  • execution model
• Simple CUDA FORTRAN example
• Compiling CUDA FORTRAN code
• Pinning memory and Asynchronous transfers
• Calling CUDA Libraries
• CUDA FORTRAN and OpenACC interoperability
• Multi-GPU programming
  • peer-to-peer
  • multiple distributed GPU and MPI

• Presentations and practicals

• CUDA FORTRAN (PGI compiler installed on HPCS systems) and PuTTY on MCS Windows

• HPCS users will have the opportunity to perform experiments directly to Wilkes
• There are few books "CUDA Fortran For Scientists and Engineers" (G. Ruetsch, M. Fatica - Morgan Kaufmann) available for students who wants to engage in long-term code porting of their FORTRAN application to CUDA FORTRAN

• One full day

• High Performance Computing: An Introduction
• High Performance Computing: Programming GPU using Open ACC
• Unix: Introduction to the Command Line Interface (Self-paced)

• High Performance Computing
• Programming and Scripting
• Scientific Computing
• UIS Systems