Thank you for sending your enquiry! One of our team members will contact you shortly.
Thank you for sending your booking! One of our team members will contact you shortly.
Duration 21 hours
Course Outline
Introduction to Biren GPU Architecture
- Overview of Biren and its primary use cases.
- Examination of hardware components: cores, memory, and compute clusters.
- Architectural comparisons with NVIDIA and AMD GPUs.
Setting Up the Biren Programming Environment
- Installation of the Biren SDK and runtime environment.
- Exploration of the toolchain and compiler architecture.
- Understanding basic project structures and build workflows.
GPU Programming with the Biren Stack
- Detailed study of thread and block organizational models.
- Management of memory resources and data transfer operations.
- Development of kernels and analysis of launch patterns.
Porting from CUDA to Biren
- Strategies for translating existing CUDA codebases.
- Mapping of common APIs and necessary code adaptations.
- Practical labs focused on code conversion and refinement.
Debugging and Profiling
- Utilizing Biren’s native debugger and profiling tools.
- Techniques for identifying performance bottlenecks.
- Optimization of memory access patterns.
Optimization Techniques
- Advanced thread scheduling and instruction pipelining strategies.
- Application of loop unrolling and efficient shared memory usage.
- Fine-tuning kernels to maximize throughput.
Case Study and Application Examples
- Real-world training of models using Biren accelerators.
- Porting and profiling vision or NLP model workloads.
- Performance benchmarking against CUDA/NVIDIA standards.
Summary and Next Steps
Requirements
- Foundational knowledge of GPU architecture and parallel processing concepts.
- Hands-on experience with CUDA, OpenCL, or comparable GPU programming frameworks.
- Familiarity with major deep learning frameworks such as PyTorch or TensorFlow.
Target Audience
- HPC developers.
- AI infrastructure engineers.
- Specialists in performance optimization.
Testimonials (2)
The extensive selection of tools presented
Miruna Buzduga - Aeronamic Eastern Europe
Course - AI Enablement Training for Engineers
Step by step training with a lot of exercises. It was like a workshop and I am very glad about that.
