The new Cadence Tensilica Vision Q7 DSP can deliver 1.82 tera operations per second (TOPS). An enhanced instruction set supports 8, 16, and 32-bit data types, and optional VFPU support for single and half precision. The Vision Q7 DSP can deliver 512 8-bit MACs for AI applications, and features enhancements including 3D DMA, compression, and 256-bit AXI interface.
The Vision Q7 DSP supports the Caffe, TensorFlow, and TensorFlowLite frameworks via the Tensilica Xtensa neural network compiler. It also supports the Android Neural Network API and more than 1,700 OpenCV-based vision library functions, and development tools and libraries are designed to achieve ISO 26262 automotive safety integrity level D certification.
Cadence’s new Vision Q7 DSP will be released in Q2, 2019.
To Learn More:
Contact: Cadence Design Systems
Headquarters: San Jose, CA, USA
Product: Tensilica Vision Q7 DSP
Key Features: 1.82 TOPS, 512 8-bit MACs for AI, support for Caffe, TensorFlow, and TensorFlowLite frameworks.
What Cadence says: View more information on the Tensilica Vision Q7 DSP.