The Intel Xeon Phi co-processor is a novel accelerator technology that possesses several appealing features, including high peak performance and the ability to use familiar programming models such as OpenMP and MPI. This course is an introduction to the Xeon Phi architecture, execution model, and programming.
Monday 2 December
Familiarity with the technology and programming model; knowledge on available tools and libraries. After the course the participants should be able to start porting their applications on Xeon Phi platforms.
Pre-requisites
The PATC course Introduction to Parallel Programming or equivalent background knowledge on MPI and OpenMP together with fluency in Fortran 95 and/or C programming. Experience of programming using basic threading is essential to fully benefit from the course. Familiarity with the following will be useful as well: CUDA, OpenCL, profiling, debugging, Intel VTune, Intel compilers, SLURM and TotalView.
Price: Free of charge
Lecturers: Olli-Pekka Lehto (CSC), Mikko Byckling (CSC), Alejandro Duran (Intel)
Monday 2 December
- 09:00 - 09:15 Course introduction
- 09:15 - 10:45 Introduction to Xeon Phi
- 10:30 - 10:45 Coffee break
- 10:45 - 12:00 Programming the Xeon Phi Coprocessor
- 12:00 - 13:00 Lunch break
- 13:00 - 13:30 CSC Cluster introduction
- 13:30 - 14:30 Exercises
- 14:30 - 14:45 Coffee break
- 14:45 - 15:45 Exercises
- 15:45 - 16:00 Wrap-up day 1
- 09:00 - 09:45 Programming Xeon Phi (MPI)
- 09:45 - 10:45 Exericises
- 10:45 - 11:00 Coffee break
- 11:00 - 12:00 Performance analysis and optimization
- 12:00 - 13:00 Lunch break
- 13:00 - 14:30 Exercises
- 14:30 - 14:45 Coffee break
- 14:45 - 15:15 Debugging
- 15:15 - 15:45 Case study
- 15:45 - 16:00 Wrap-up day 2
Familiarity with the technology and programming model; knowledge on available tools and libraries. After the course the participants should be able to start porting their applications on Xeon Phi platforms.
Pre-requisites
The PATC course Introduction to Parallel Programming or equivalent background knowledge on MPI and OpenMP together with fluency in Fortran 95 and/or C programming. Experience of programming using basic threading is essential to fully benefit from the course. Familiarity with the following will be useful as well: CUDA, OpenCL, profiling, debugging, Intel VTune, Intel compilers, SLURM and TotalView.
Price: Free of charge
Lecturers: Olli-Pekka Lehto (CSC), Mikko Byckling (CSC), Alejandro Duran (Intel)