HPC Master’s Degree Repository

Welcome to my repository dedicated to tasks, assignments, and investigations related to my Master’s Degree in High-Performance Computing (HPC). This repository serves as a structured collection of source code, scripts, tools, and reference materials developed or studied throughout the course.

---

🗂️ Repository Structure

This repository is organized into the following main directories:

1. cuda/
   Contains assignments and projects completed as part of the coursework. Examples include:

   • CUDA Applications: Parallelized algorithms using NVIDIA CUDA.
   • OpenMP Implementations: Shared memory parallelism in C/C++.
   • MPI Programs: Distributed memory solutions for clusters.

2. cloud/
   Scripts and configurations for setting up and managing HPC workloads on cloud platforms, particularly AWS. This includes:

   • Job submission scripts for AWS Batch.
   • Auto-scaling configurations for HPC clusters.
   • Optimization for cloud-based compute nodes.

3. hpc_tools/
   Tools and utilities for profiling, debugging, and analyzing performance:

   • Usage of NVIDIA Nsight, gprof, Valgrind, and others.
   • Performance tuning and benchmarking tools.
   • Custom scripts for automating performance analysis workflows.

4. mpi/
   Documentation, code samples, and notes based on:

   • Books: Summaries and examples from key HPC textbooks.
   • Papers: Reproducing experiments or methodologies from academic research.
   • Trainings: Exercises and examples from HPC-related training courses.

5. omp/
   Documentation, code samples, and notes based on:

   • Books: Summaries and examples from key HPC textbooks.
   • Papers: Reproducing experiments or methodologies from academic research.
   • Trainings: Exercises and examples from HPC-related training courses.

6. opencl/
   Documentation, code samples, and notes based on:

   • Books: Summaries and examples from key HPC textbooks.
   • Papers: Reproducing experiments or methodologies from academic research.
   • Trainings: Exercises and examples from HPC-related training courses.

---

💡 Objectives

This repository aims to:

• Demonstrate hands-on knowledge in programming for parallel and distributed systems.
• Document the learning journey in HPC concepts and their practical applications.
• Facilitate sharing of insights, examples, and best practices in HPC development.

---

🔧 Tools & Technologies

The repository includes work involving the following tools and technologies:

• Programming Languages:
  Primarily C and C++ with extensions for CUDA, OpenMP, and MPI.

• HPC Tools:

  • Profilers: NVIDIA Nsight, gprof.
  • Debuggers: Valgrind, CUDA Debugger.
  • Performance Analyzers: Perf, VTune.

• Cloud Platforms:

  • AWS for on-demand HPC resources.
  • Scripts for auto-scaling and cost-efficient usage.

---

🚀 How to Use

1. Clone the repository:

   git clone https://www.usc.gal/en/studies/masters/engineering-and-architecture/master-high-performance-computing-online.git
   
   

2. Navigate to the appropriate folder for the task or topic you are interested in:

   cd hpc-masters/assignments
   

3. Follow the instructions or README files within each folder to build, execute, or explore the materials.

📚 Resources & References

The following resources have been instrumental in this journey:

Books: "Parallel and High Performance Computing; C++ Concurrency in Action; " and others.
Trainings: NVIDIA Deep Learning Institute, AWS HPC Workshops; HLRS
Papers: ACM Research articles on performance analysis, scaling HPC applications, and optimization strategies.