Leuven, Belgium – September 29, 2022 Battery powered Extended Reality (XR) devices need to sense and analyze their environments during the response time of the human senses to provide instant visual and auditory feedback towards the user. And do it with ultra-low power consumption and ultra-low latency. The easyics solution meets these requirements with integrated nearbAI semiconductor IP cores and unique L3-Optimizer technology for low power + low latency + low area balance. Digital IP cores are optimized for integration into ASICs, ASSPs, and SoCs within XR hardware and are available as of today.
The growth of edge computing and artificial intelligence has increased the need for high-performance neural network inference engines on high-end devices. Gartner predicts that by 2027, deep learning will be included in more than 65% of edge use cases. XR edge devices in the form of virtual reality (VR) headsets and augmented reality (AR) glasses are expected to experience double-digit growth and exceed 50 million units by 2026, according to IDC. XR devices will use increasingly sophisticated artificial intelligence to sense and analyze their environments locally, as part of the metaverse’s distributed computing landscape.
nearbAI IP cores are highly efficient, configurable neural network inference engines for a wide and growing range of AI models and use cases optimized for power consumption, performance, and area. nearbAI works on functions such as scene segmentation and reconstruction, and face and face detection and recognition, which are central to consumer and enterprise XR applications.
“nearbAI is unique, offering best-in-class cutting-edge AI inference solutions for XR devices. It is based on more than three decades of DSP design and image sensor chips. Our mission is to provide customers with IP addresses that perfectly match their specifications, deliver outstanding performance, It can be easily combined by ejecting the tape”, said Ramses Valfkins, CSO and Managing Director of easyics.
Ideal for mobile XR processor chips, it also fits more compact AI chips near the sensors:
- Maximum optimization
- Special L3 Optimizer: Low Power + Low Latency + Low Area Balance
- Scaled and tuned according to the application, with the possibility of upgrading the field
- Record-breaking MAC usage up to 95%
- Wide configurable number of MACs: 16 to 4096
- Configurable MAC precision, parameter independent and data quantization, 4 to 16 bits, single bit precision
- Configurable internal memory bank sizes and bus widths
- low energy
- As low as less than 10mW
- Ultra fast response time
- Face detection in 2 milliseconds
- zero latency shift
- Multiplexing continuous transmission between multiple neural networks on the same nearbAI core, saving the silicon region
- Support for any sensor
- RGB, ToF, NIR, Thermal IR, LiDAR, Hyperspectral, Stereo, Ultrasound, Audio
“ISEX is committed to bringing the latest technologies and solutions to market. With nearbAI, together with our partners and customers, we push the latest AI innovations to the limits,” Mr. Valkins added.
To learn more about nearbAI for XR devices and experience a live demo, visit the Isis booth at Augmented Enterprise Summit (AES) 2022 in San Diego, October 18-20, Booth 802 (www.augmentedenterprisesummit.com), or visit www.nearbai.com
easyics is a market leader in digital chip design and IP licensing for semiconductors. With more than 30 years of design excellence in embedded solutions, ESIX delivers unique efficiencies and development platforms that lead to true, reliable, and optimized logic for the first time. easics supports leading OEMs and semiconductor companies with custom designs and customizable IP blocks for intelligent embedded processing systems achievable in custom ASICs/ASSPs/SoCs and in FPGAs.
For more information, please visit the website www.easics.com .