Industry’s first multi-modal, software-centric edge AI platform supports any edge AI model from CNNs to multi-modal GenAI and everything in between with scalable performance per watt
SiMa.ai, the software-centric, embedded edge machine learning system-on-chip (MLSoC) company, announced MLSoC™ Modalix, the industry’s first multi-modal edge AI product family. SiMa.ai MLSoC Modalix supports CNNs, Transformers, LLMs, LMMs and Gen AI at the edge, and delivers industry leading performance – more than 10X the performance per watt of alternatives.
The rise of generative AI is changing the way humans and machines work together. The next wave of the AI technology revolution will advance multi-modal machines with the ability to understand and process multiple forms of inputs across text, image, audio and visual. This shift will ripple across every industry, from agriculture and logistics, to medicine, defense, transportation and more. SiMa.ai is the breakthrough multi-modal edge AI platform that organizations can now rely on for the ultra-responsive, power-efficient, reliable and secure insights their critical innovations demand – in any modality they prefer.
MLSoC Modalix will enable developers to push the envelope on performance with a small power envelope and physical footprint across a wide breadth of applications. They will be able to run end to end GenAI application-pipelines on just a single chip while delivering measurable contributions to business top and bottom lines.
SiMa.ai MLSoC Modalix is the second generation of the successful, commercially deployed first generation MLSoC. MLSoC Modalix is offered in 25 (Modalix 25 or “M25”), 50 (Modalix 50 or “M50”), 100 (Modalix 100 or “M100”) and 200 (Modalix 200 or “M200”) TOPS configurations, in multiple form factors, and is purpose-built to provide effortless deployment of Generative AI for the embedded edge ML market. Fully software compatible with first generation MLSoC, the MLSoC Modalix product family was designed to enable the capability to run DNNs, as well as advanced Transformer models, including LLMs, LMMs and Generative AI. Samples of MLSoC Modalix will be available to customers in Q4 of 2024.
“SiMa.ai is leading the edge AI market and building increasing momentum as customers around the world continue to adopt our (first generation) MLSoC. We are excited to introduce the MLSoC Modalix family to give customers even greater choice and flexibility to embrace multi-modal AI,” said Krishna Rangasayee, Founder and CEO at SiMa.ai. “With the introduction of MLSoC Modalix, SiMa.ai’s ONE Platform for Edge AI now provides coverage from CNNs to Gen AI and everything in between with industry leading performance and power efficiency.”
SiMa.ai Broadens ONE Platform for Edge AI with New MLSoC Modalix Family
SiMa.ai ONE Platform for Edge AI is a software-centric framework offering compatibility across the entire MLSoC platform family to provide a seamless experience for upgrades, transitions and mix-and-match, minimizing TCO for adopters of SiMa.ai at the edge.
First generation SiMa.ai MLSoC is a market leader in performance and power efficiency, validated most recently by MLCommons® MLPerf Inference benchmark (4.0) results in March 2024. First generation SiMa.ai MLSoC will continue to support CNN-based models across Industrial Inspection, Retail, Aerospace and Defense, and Smart Vision (Ethernet-based cameras only).
Also Read: Palantir Named a Leader in AI/ML Platforms
Customers leveraging MLSoC Modalix will be able to accelerate deep learning, transformer, and Generative AI models – pushing the envelope on models like Llama2 7B to beyond 10 tokens per second. With an on-chip application processor, integrated ISP, vision and digital signal processors, and innovations like BF16 in the hardware, MLSoC Modalix will accelerate all known CNN and transformer models: language, vision and speech, delivering inference-accuracy close to fp32 at significantly lower cost and power.
Generative AI applications running on MLSoC Modalix will:
- Benefit from the SiMa.ai Palette software patented Compiler with layered direct proactive data prefetch, ensuring each data layer is always available, on time, when it is needed;
- Leverage the improved Machine Learning Accelerator to run DNNs and Large Multi-Modal Models at high performance without sacrificing power.
In addition new hardware components of MLSoC Modalix include:
- An integrated ISP module supporting 4 x 4 lanes of MIPI CSI 2, allowing the chip to interface with MIPI cameras as well as Ethernet-based cameras via 4 x 10G Ethernet ports on-chip;
- Eight lanes of PCIe Gen 5 facilitating quicker ingress and egress of data to and from the chip;
- 8 Arm® Cortex®-A65 dual-threaded CPUs, further extending SiMa.ai’s reach to customers with specialized use cases through Arm’s unrivaled and diverse software partner ecosystem;
- TSMC’s N6 technology.
Industry Validation for SiMa.ai
“The launch of the SiMa.ai ONE Platform with the MLSoC Modalix family has the potential to be a great enabler for Elementary’s applications,” said Arye Barnehama, CEO of Elementary. “Elementary is at the forefront of deploying AI-based vision inspection systems for global manufacturers dedicated to leading-edge innovation. The Modalix family’s high performance and energy efficiency, coupled with the ONE Platform approach, aligns with our commitment to delivering modular, uncompromising solutions to both our partners and end users. As Elementary continues to scale in the industrial sector, the SiMa.ai MLSoC Modalix product family has the potential to power solutions that are integral to the high-speed production processes of global enterprises.”
“The SiMa.ai MLSoC Modalix family has the ability to accelerate a wide variety of architectures ranging from CNNs to LMMs and in turn make multi-modal edge AI on power-constrained platforms a reality. Hayden AI is excited at the prospect of evaluating the SiMa.ai Modalix family for potential inclusion in our future products,” said Vaibhav Ghadiok, CTO, Hayden AI.
“The availability of SiMa.ai MLSoC Modalix product family is a game changer for us and our edge AI customers. Advantech has been building and deploying industrial PCs, smart cameras, and edge gateways for healthcare, industrial automation, and smart cities for many years,” said Emily Teng, Associate Director of Product Management, Advantech. “Rich peripherals including MIPI and advanced ISP combined with the high-performance and low-power envelope of the SiMa.ai MLSoC Modalix family will enable a wide range of smart devices at the edge. SiMa.ai will allow us to adapt and scale to embrace new modalities and innovation inherent in GenAI.”
“The availability of SiMa.ai ONE Platform with the MLSoC Modalix family is a potential game changer for Edge-AI applications that LIPS Corporation has been building and deploying to scale multi-camera 3D sensing solutions in smart factory and smart vision projects,” said Luke Liu, CEO of LIPS Corporation. “The high-performance and low-power envelope of the MLSoC Modalix family, combined with their ONE Platform approach, can enable us to overcome the performance and bandwidth limitations often associated with multi-camera 3D sensing projects. SiMa.ai MLSoC Modalix can enable us to adapt and scale new AI modalities and innovations in real-time pose detection, precise spatial mapping, and advanced AI detection across various industrial and automation settings, ultimately accelerating our ability to deliver smarter, more efficient, and safer solutions for our clients.”
“In the age of AI, hardware and software capabilities need to advance at a faster pace than ever before,” said Paul Williamson, SVP and GM, IoT Line of Business, Arm. “By building solutions like MLSoC Modalix on the high performance, efficient Arm platform, partners like SiMa.ai can access a powerful software ecosystem to create the next generation of secure and innovative user experiences across applications ranging from retail to industrial.”
“TSMC has long-partnered with industry innovators like SiMa.ai, helping them achieve groundbreaking next-generation semiconductor designs with our industry-leading technologies and manufacturing excellence,” said Lucas Tsai, Senior Director of Market Development and Emerging Business Management, TSMC North America. “We are excited to continue our collaboration in accelerating SiMa.ai’s powerful and energy-efficient chip innovation to meet the rapidly growing demand for GenAI at the edge.”
“Synopsys continues to help companies at all levels accelerate and scale their advanced AI chip development process,” said Ravi Subramanian, GM of the Systems Design Group at Synopsys. “By leveraging Synopsys’ leading AI-driven EDA suite, broad interface, security, and processor IP portfolio, architecture design solution and emulation systems, SiMa.ai is speeding development cycles and unlocking step-function gains in performance and power for their next-generation MLSoC Modalix family. Together, we are pushing the boundaries of what is possible with GenAI applications, empowering engineers to be more productive and proficient, and shaping the future of AI at the edge.”
Source: Businesswire
