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 14 hours
Course Outline
Overview of CANN Optimization Features
- How CANN manages inference performance
- Optimization objectives for edge and embedded AI environments
- Insights into AI Core usage and memory distribution
Leveraging Graph Engine for Analysis
- Foundations of the Graph Engine and its execution pipeline
- Visualizing operator graphs and runtime indicators
- Adjusting computational graphs to achieve optimization
Profiling Utilities and Performance Indicators
- Utilizing the CANN Profiling Tool (profiler) for workload assessment
- Evaluating kernel execution times and identifying bottlenecks
- Memory access analysis and tiling methodologies
Developing Custom Operators with TIK
- Overview of TIK and its operator programming framework
- Building a custom operator using the TIK DSL
- Conducting tests and performance benchmarks for operators
Advanced Operator Tuning with TVM
- Introduction to TVM integration within CANN
- Auto-tuning techniques for computational graphs
- Determining when and how to alternate between TVM and TIK
Memory Optimization Strategies
- Controlling memory layout and buffer positioning
- Methods to lower on-chip memory usage
- Best practices for asynchronous execution and resource reuse
Practical Deployment and Case Studies
- Case study: Tuning performance for smart city camera systems
- Case study: Optimizing the inference stack for autonomous vehicles
- Guidelines for iterative profiling and ongoing enhancement
Requirements
- A robust command of deep learning model architectures and training processes
- Hands-on experience deploying models via CANN, TensorFlow, or PyTorch
- Proficiency in Linux CLI, shell scripting, and Python programming
Target Audience
- AI performance engineers
- Specialists in inference optimization
- Developers focused on edge AI or real-time systems
