Cadence boosts vision and AI performance with Tensilica Vision Q6 DSP IP
Using a fast processor architecture, the fifth-generation Cadence Tensilica Vision Q6 DSP targets embedded vision in the smartphone, surveillance camera, automotive, AR/VR, drone and robotics markets. The fifth-generation Vision Q6 DSP offers 1.5X greater vision and AI performance than its predecessor, the Vision P6 DSP, and 1.25X better power efficiency at the Vision P6 DSP’s peak performance.
It has a deeper, 13-stage processor pipeline and system architecture designed for use with large local memories. This enables the Vision Q6 DSP to achieve 1.5GHz peak frequency and 1GHz typical frequency at 16nm, in the same floorplan area as the Vision P6 DSP.
An enhanced DSP instruction set results in up to 20 per cent fewer cycles than the Vision P6 DSP for embedded vision applications/kernels such as Optical Flow, Transpose and warpAffine, and commonly used filters such as Median and Sobel
Doubling system data bandwidth with separate master/slave AXI interfaces for data/instructions and multi-channel division multiple access (DMA) alleviates memory bandwidth challenges in vision and AI applications, says Cadence. They also reduce latency and overhead associated with task switching and DMA set-up
The DSP is backward compatible with the Vision P6 DSP, to preserve software investment for an easy migration.
The Vision Q6 DSP supports AI applications developed in the Caffe, TensorFlow and TensorFlowLite frameworks through the Tensilica Xtensa Neural Network Compiler (XNNC). The XNNC maps neural networks into executable and highly optimised high-performance code for the Vision Q6 DSP, leveraging optimised neural network library functions. The Vision Q6 DSP also supports the Android Neural Network (ANN) API for on-device AI acceleration in Android-powered devices. The software environment also features complete and optimised support for more than 1,500 OpenCV-based vision and OpenVX library functions, enabling fast, high-level migration of existing vision applications.
The Vision Q6 DSP is designed for general-purpose embedded vision and on-device AI applications requiring performance ranging from 200 to 400GMAC per second. Peak performance is 384GMAC per second, making the Vision Q6 DSP suitable for high-performance systems and applications, says Cadence. The Vision Q6 DSP can be paired with the Vision C5 DSP for applications requiring greater than 384 GMAC per second AI performance.
Picture credit: Drone by DJI