The Ensemble E4 (MCU) uses the Arm dual Cortex-55 core and the Ensemble E6 and E8 fusion processors are based on the Arm Cortex-32 core and dual Cortex-32 cores respectively. Alif said it is the first silicon provider to use an Arm Ethos-U85 NPU (neural processing unit) which supports transformer-based machine learning networks, reported by Electronics Weekly earlier this year.
“AI is changing the expectation of edge devices and it is changing very fast,” said Reza Kazerounian, president of Alif Semiconductor, explaining that these three devices are the company’s way for getting ready for these changes. They are designed for scalable processing, migrating from single to multi-core processing using the same architecture and software.
All use proprietary Autonomous Intelligent Power Management technology, consuming 1.3μA in stop mode and down to 27μA/MHz in dynamic consumption mode. Different on-chip (NVRAM and SRAM) and external memory interface options are available, depending on the model.
The second generation Ensemble devices support up to two MIPI-CSI image sensors. Benchmark tests show power efficient object detection in less than 2ms, and image classification in less than 8ms. The devices also include a hardware-accelerated image signal processor (ISP) pipeline that operates at up to 60fps at 2MP resolution.
For security, all have hardware-based root-of-trust with unique device ID, secure key generation and storage, secure certificate storage, secure boot, in-factory provisioned private keys and a selection of crypto accelerators (AES, ECC, SHA, RSA and NIST-compliant TRNG.
Typical applications are in consumer and industrial products, for example human-to-computer interfacing, wearable devices, healthcare and diagnostics, robotics, transportation, toys, smart home devices and smart city equipment.
Development kits, based on the Ensemble E8, and therefore suitable for evaluated all three devices, are also available.
