Displacing alternatives with purpose-built technology
SiMa.ai, the machine learning company enabling effortless ML deployment and scaling at the embedded edge, today announced availability of two new PCIe-based production boards that scales embedded edge ML deployments for key customers. The availability of these two new commercially deployable board-level products demonstrates the commitment of SiMa.ai’s mission to simplify ML scalability at the embedded edge. The company also announced its Palette™ software that provides a pushbutton experience for developing complete end-to-end ML applications targeting the heterogeneous SiMa.ai Machine Learning SoC (MLSoC™) platform.
The newly unveiled SiMa.ai PCI Express Half-height Half-length (PCIe HHHL) and Dual M.2 production boards are purpose-built with the MLSoC platform silicon to require less power for utilizing ML at the edge. The efficiency of the MLSoC architecture provides the ability to meet the power constraints of the smallest embedded edge form factors. SiMa.ai’s 10x performance advantage provides headroom to continually innovate after deployment with any new algorithms and networks.
“We’re excited to bring these new form factor boards, programmed with our Palette software, to market for our customers because they address a growing need for a combined complete software and hardware solution within the developer community,” said Krishna Rangasayee, CEO and Founder, SiMa.ai. “Developers being empowered to not only develop but to deploy any ML vision application with 10x better performance will be a game changer for our ever-expanding list of customers.”
The PCIe HHHL and Dual M.2 are versatile production boards that use the SiMa.ai MLSoC Platform, providing a choice for customers to deploy quickly without waiting for internal board development cycles to enter production. Customers can use these board designs to accelerate deployment and use the design to develop their own custom form factors as needed, while quickly deploying ML at the edge.
The SiMa.ai MLSoC device offers heterogeneous cores for processing anycomputer vision ML workload. These heterogeneous compute elements include quad Arm A65 cores, a Video encoder/decoder that supports the H.264 standard, a Machine Learning Accelerator (MLA) block that provides up to 50 TOPS for ML acceleration along with a Computer Vision Processor (CVP) to any ML computational needs for any framework. The boards’ standard form factor and proven design eliminates the need for new or customized hardware by customers. Commercial versions are available with pricing in 10K unit quantities at $599 for Dual M.2 and $749 for the PCIe HHHL. Industrial temperature grade versions are planned.