To increase awareness of accelerated computing, we are holding this Bootcamp. Registration is recommended for individuals interested in learning accelerated programming and making use of accelerated applications so that results can be obtained more quickly. JNCASR has many GPUs installed in its NSM cluster, and users are welcome to utilise them.

Agenda:
Day 1: 16th January ( 10:00 AM – 5:00 PM )
• Lecture : Welcome Address and Overview of Workshop
• Lecture : C-DAC HPC System Software Stack for PARAM Yukti

• Lecture : Intro to Linux and Shell scripting

• Lecture : User Management, HPC System Monitoring, Ticketing Support

• Lecture : Application porting demo using SPACK on PARAM Yukti

• Lecture : SLURM (Workload manager)

• Lab Session : SLURM job submission/Ticketing/Support • Q and A Session

Day 2: 17th January ( 10:00 AM – 5:00 PM )

• Lecture : Shared Memory Parallelism with OpenMP

• Lecture : MPI Point to Point Communication

• Lecture : MPI Collective communication

• Lab Session : Shared Memory Parallelism with open MPI

• Lab Session : Distributed Memory Programming with MPI

• Q&A

Day 3: 18th January ( 10:00 AM – 5:00 PM )
• Lecture : Introduction to Accelerated Computing

• Lab Session : Using GPU on PARAM Yukti

• Lab Session : MD Case studies: GROMACS/LAMMPS (CPU vs GPU)

• Q & A Session

Day 4: 19th January ( 10:00 AM – 5:00 PM )
• Lecture : Introduction to Open ACC

• Lab Session : Using Parallel Directive

• Lab Session : Using Data Directive

• Lab Session : Optimization

Day 5: 20th January ( 10:00 AM – 5:00 PM )
• Lecture : Introduction to CUDA

• Lab Session : CUDA Memory Model

• Lab Session : CUDA Programming Model

• User Code Case studies

• Q & A Session