- Semiconductor BusinessHOME
- Products and Services of Macnica,Inc.
-
technical information
-
Events and Seminars
- Handling Manufacturer
- Support
- Inquiry
- Click here to purchase products
- Semiconductor business e-mail magazine registration
Product Summary
The MLSoC™ provided by SiMa.ai is a System-on-Chip (SoC) specialized for inference processing designed for embedded applications. Manufactured using a 16nm process, this product is capable of completing a series of processes, from image and video input to pre-processing, inference, post-processing, and output, all on a single chip.
MLSoC™ combines industry-leading low power consumption performance with highly efficient inference processing capabilities, making it ideal hardware for Physical AI applications that require real-time performance and low power consumption.
In addition, SiMa.ai offers "Palette™," software that works with MLSoC™, significantly improving developer productivity through GUI-based AI pipeline construction and compatibility with existing AI frameworks (TensorFlow, PyTorch, etc.).
Product Specifications
MLSoC™ is an SoC that enables low-power, high-performance edge AI processing, achieving image processing and inference on a single chip.
| TOPS | 50 |
| Machine Learning Accelerator (MLA) |
INT8 / INT16 |
| Application Processor |
Arm A65 x 4 dual-threaded 1.15GHz; 15K DMIPS |
| Vision Processor | Synopsys ARC EV-74 800 MHz |
| Video Decoder | H.264/265 4KP60 x 1 |
| Video Encoder | H.264 4KP30 x 1 |
| PCIe | Gen 4 x 8 lanes endpoint only |
| Ethernet | 1GbE x 4 |
| On-chip Memory | 4 MB |
| DRAM | 32-bit LPDDR4 x 4 |
| Manufacturing Process Node | TSMC 16nm |
| Model Types | CNNs, Transformers |
Product boards
MLSoC™ is available not only as an SoC, but also as PCIe HHHL and PCIe Dual M.2 boards.
PCIe HHHL
*For detailed specifications, please see the Product Brief.
PCIe Dual M.2
*For detailed specifications, please see the Product Brief.
Evaluation Kit - MLSoC™ DevKit -
The MLSoC™ DevKit is a high-performance development platform designed for computer vision and machine learning applications at the edge, enabling efficient model evaluation and prototyping, while delivering real-world performance with low latency and industry-leading power consumption.
| Ethernet Ports (eth0, eth1) | RJ45 GigE interface for remote connectivity when using the DevKit standalone |
| PCI-Express 4.0 x4 Interface | High-speed PCIe interface used for integration with the host machine |
| Operation Mode Switch |
All switches ON: Standalone mode Only switch 1 is OFF: PCIe mode |
| 12V Power (J8) | Powers the DevKit in standalone mode (12V power adapter included) |
| I2C (J1) | I2C communication is possible for control or data communication |
| GPIO Header (J2) | Provides general-purpose input/output (GPIO) pins for custom applications, such as controlling external devices and sensors |
| COM Port to Host (J4) | Used for serial connection to the host machine (USB serial cable included) |
| Reset Button | Only available in standalone mode |
| Debug COM Port (J5) | Debug COM port for T-root access and device flashing |
| Run/Program Switch |
ON: Normal operation mode OFF: Programming mode for firmware updates or setting changes |
| SD Card Slot | A slot is provided on the back of the MLSoC™ DevKit |
Inquiry
Please feel free to contact us with any questions about our products, technical inquiries, sample requests, or estimates.