Course name: Cuda to Sycl migration: An example based approach
- CUDA to SYCL Migration Introduction
- SYCLomatic Tool Introduction and Usage
- Migration Workflow Overview
10:30 AM – 11:00 AM Hands-on Exercise: SYCLMigration – Simple VectorAdd
- Learn how to migrate a simple single-source CUDA code to SYCL
- Follow along with a Jupyter notebook example
11:00 AM – 11:30 AM Hands-on Exercise: SYCLMigration – Sorting Networks
- Learn how to migrate a CUDA project with multiple source files that uses Makefile for the project
- Use a Jupyter notebook for guidance and practice
11:30 AM – 11:45 AM Coffee Break
11:45 AM – 12:15 PM Hands-on Exercise: SYCLMigration – Jacobi Iterative
- Learn how to migrate a CUDA project that uses CUDA features to access the GPU hardware like Shared Local Memory, warps, and atomics in kernel code
- Follow a Jupyter notebook example
12:15 PM – 12:30 PM Discussion: Challenges and Solutions in Migration
- Discuss common challenges faced during the migration process
- Share tips and solutions to overcome these challenges
12:30 PM – 1:30 PM Lunch Break
1:30 PM – 2:15 PM Hands-on Exercise: SYCLMigration – Matrix Multiplication with CuBlas Library
- Learn how to migrate a CUDA project that uses the cuBLAS library
- Use a Jupyter notebook for guidance
2:15 PM – 3:00 PM Hands-on Exercise: SYCLMigration – OceanFFT with oneMKL FFT Library
- Learn how to migrate a CUDA project that uses FFT from the oneMKL library
- Follow along with a Jupyter notebook example
3:00 PM – 3:15 PM Coffee Break
3:15 PM – 3:45 PM Hands-on Exercise: SYCLMigration – SimpleCUDAGraphs
- Learn how to migrate a CUDA Graph feature project manually using SYCL Taskflow and SYCL Graphs
- Use a Jupyter notebook for guidance
3:45 PM – 4:15 PM Hands-on Exercise: SYCLMigration – MonteCarloMultiGPU
- Learn how to migrate a CUDA project that uses RNG from the oneMKL library
- Follow along with a Jupyter notebook example
4:15 PM – 4:30 PM Wrap-Up and Review
- Recap key concepts and takeaways
- Address any remaining questions
- Feedback and discussion
Date:- 16 December, 2024
Time:– 10.00 – 16.30
Venue:- Room: ASL 5.05 – SALC 5 (Access through SHERFIELD BUILDING)