Get in Touch
 Duration 21 hours

Course Outline

Biren GPU Architecture Fundamentals

  • Overview of Biren and its primary use cases
  • Hardware configuration: cores, memory structure, and compute clusters
  • Comparative analysis with NVIDIA and AMD GPU architectures

Configuring the Biren Programming Environment

  • Installation of the Biren SDK and runtime components
  • Exploring the toolchain and compiler architecture
  • Structuring basic projects and managing the build process

GPU Programming Using the Biren Stack

  • Concepts of thread and block organization
  • Managing memory and handling data transfers
  • Developing kernels and establishing launch patterns

Migrating from CUDA to Biren

  • Strategies for translating CUDA codebases
  • Mapping common APIs and adapting to new interfaces
  • Hands-on labs for code conversion and practice

Debugging and Profiling Processes

  • Utilizing Biren’s integrated debugger and profiler tools
  • Locating and analyzing performance bottlenecks
  • Optimizing memory access patterns for efficiency

Advanced Optimization Techniques

  • Managing thread scheduling and instruction pipelining
  • Implementing loop unrolling and leveraging shared memory
  • Advanced kernel adjustments to maximize throughput

Case Studies and Real-World Applications

  • Executing model training with Biren accelerators
  • Porting and profiling computer vision or NLP models
  • Benchmarking performance against CUDA/NVIDIA stacks

Conclusion and Future Pathways

Requirements

  • A solid understanding of GPU architecture and parallel processing concepts.
  • Prior experience with CUDA, OpenCL, or comparable GPU programming frameworks.
  • Familiarity with deep learning libraries such as PyTorch or TensorFlow.

Target Audience

  • Developers specializing in HPC.
  • Engineers working in AI infrastructure.
  • Specialists focused on performance optimization.

Number of participants


Price per participant

Testimonials (2)

Upcoming Courses

Related Categories