I805 utilizes a 4-stage sequential pipeline, and is equipped with a vector computing engine oriented to applications such as AI and DSP. It is designed with low-latency tightly coupled memory (TCM) to ensure excellent data throughput efficiency. It is suitable for application fields with certain requirements on computing power, such as audio codec, voice processing, and lightweight deep learning networks.
AIoT processor with vector computing engine
Overview
Key Features
- Instruction set: T-Head ISA (32-bit/16-bit variable-length instruction set);
- Pipeline: 4-stage sequential pipeline;
- General register: 32 32-bit GPRs; 16 128-bit VGPRs;
- Cache: I-Cache: 8 KB/16 KB/32 KB/64 KB (size options); D-Cache: 8 KB/16 KB/32 KB/64 KB (size options);
- Tightly-coupled memory (TCM): I-TCM: 4 KB to 1 MB (size options); D-TCM: 4 KB to 1 MB (size options);
- Tightly-coupled memory slave interface: Independent TCM bus slave interface;
- Bus interface: Dual bus (system bus + peripheral bus);
- Memory protection: 0 to 8 optional protection zones;
- Scalar computing engine: 32-bit operation width;
- Vector computing engine: 128-bit operation width;
- Tight coupling IP: Interrupt controller and timer;
- Floating point engine: Optional single-precision floating point unit;
- Vector calculation engine: Improves computing parallelism, and accelerates application scenarios like AI;
- Low-latency tightly-coupled memory: Expands memory bandwidth, and adapts to data-intensive computing scenarios;
- High-performance unaligned memory access: Accelerates unaligned memory access, and adapts to DSP applications.
Block Diagram
Applications
- Speech Recognition;
- Smart Home Appliances.
Technical Specifications
Related IPs
- High-performance 64-bit RISC-V architecture multi-core processor with AI vector acceleration engine
- High-performance 32-bit multi-core processor with AI acceleration engine
- Low power 32-bit processor with lightweight computing power
- 2.5D Multi-Core Raster & Vector Graphics Processor for low-power SoCs with Microcontroller
- High-performance AI dataflow processor with scalable vector compute capabilities
- Highest code density, Low Power 32-bit Processor with optional DSP