Polish embedded systems company Grinn has recently introduced ReneSOM-V2H, a tiny vision AI SoM built around the Renesas RZ/V2H vision AI processor. Measuring just 42.6 x 37 mm, Grinn claims it is the world’s smallest module based on this specific Renesas MPU, and targets space-constrained Edge AI applications such as smart cameras, robotics, and industrial automation.
The RZ/V2H SoC features a heterogeneous architecture with 4x Cortex-A55 cores, 2x Cortex-R8 cores, and 1x Cortex-M33 core, along with a DRP-AI3 accelerator with up to 8 TOPS. It supports LPDDR4 memory and eMMC storage, along with various connectivity options, including PCIe Gen3 (4-lane), USB 3.2, USB 2.0, and Gigabit Ethernet. It also provides four MIPI-CSI camera inputs and a MIPI-DSI display output for vision applications.
Grinn ReneSOM-V2H specifications:
- SoC – Renesas RZ/V2H
- CPU/MCU cores
- 4x Arm Cortex-A55 cores up to 1.8 GHz
- 2x Cortex-R8 real-time cores up to 800 MHz
- Arm Cortex-M33 microcontroller core up to 200 MHz for system management
- GPU – Arm Mali-G31 GPU
- ISP – OpenCV Accelerator, ISP (optional Arm Mali-C55)
- NPU
- DRP-AI3 up to 8 TOPS (INT8) or 80 TOPS (Sparse)
- DRP dynamically reconfigurable processor (STP4)
- CPU/MCU cores
- System Memory – Up to 8GB supported by SoC (exact memory configuration on module not specified)
- Storage – eMMC flash (exact configuration not specified)
- LGA balls
- Display Interface – MIPI DSI
- Camera Inputs – 4x 4-lane MIPI CSI-2 interfaces
- Networking – Gigabit Ethernet interface
- USB
- 1x USB 3.2 interface
- 1x USB 2.0 interface
- Expansion
- PCIe Gen3 (4-lane)
- 6x CAN FD, UART, I2C, SPI, ADC (should have support for this, but not confirmed by the company)
- Power Supply – 5V single supply
- Dimensions – 42.6 x 37 mm (LGA design with 260-pin SO-DIMM compatibility, whatever that means)
- Operating Temperature – −40°C to +85°C (industrial range)
Looking at the specifications, it seems that the module can handle up to four cameras with high-speed data throughput via PCIe Gen3 and USB 3.2. Which makes it very similar to a NVIDIA Jetson Orin Nano. Though the raw TOPS might seem lower (8 vs 20/40), the DRP-AI3 architecture is often more efficient for specific vision pipelines without requiring a fan.
The company does not mention anything about software. However, considering the module is based on a Renesas RZ/V2H SoC, it should have support for the full Renesas software ecosystem, including a Linux BSP based on Yocto, DRP-AI drivers, and the DRP-AI TVM backend for TensorFlow models. Additional support includes Flexible Software Package (FSP) components for Cortex-R8 and Cortex-M33 real-time cores, multi-OS or bare-metal development options, and Edge Impulse integration for DRP-AI-accelerated models.
Grinn suggests that using the ReneSOM-V2H can shorten development cycles by up to 12 months compared to a chip-down design, as it handles the complexities of high-speed routing for the RZ/V2H’s 1368-pin BGA package. The Grinn ReneSOM-V2H is currently sampling, and the company did not provide pricing information. More details can be found on the product page and the official announcement.
Debashis Das is a technical content writer and embedded engineer with over five years of experience in the industry. With expertise in Embedded C, PCB Design, and SEO optimization, he effectively blends difficult technical topics with clear communication
Support CNX Software! Donate via cryptocurrencies, become a Patron on Patreon, or purchase goods on Amazon or Aliexpress. We also use affiliate links in articles to earn commissions if you make a purchase after clicking on those links.
