AI processor delivers NPU performance up to 2.4 TOPs

AI processor delivers NPU performance up to 2.4 TOPs

New Products |
By Wisse Hettinga

RK3399Pro boasts an NPU (Neural Network Processing Unit) computing performance that reaches 2.4 TOPs – a performance that is 150 percent higher than similar NPU processors, while keeping power consumption to under 1 percent, compared to other AI processors using a GPU as the AI computing unit.

The first AI processor from Rockchip, the RK3399Pro adopts a CPU+GPU+NPU hardware structure and delivers improved computing performance of typical deep neural network Inception V3, ResNet34 and VGG16 models of nearly hundred times, compared to traditional processors.

The AI processor adopts a big.LITTLE CPU, Dual-core Cortex-A72 and Quad-core Cortex-A53. The Quad-core ARM high-end Mali-T860 GPU which integrates more bandwidth compression techniques drives overall performance.

RK3399Pro AI processor supports dual Type-C interfaces, dual ISPs (each channel supporting a maximum 13 mega pixels); 4096 x 2160 display output; and  an 8-channel digital microphone array input. With respect to software, RK3399Pro supports numerous APIs, including OpenGL ES 1.x/2.x/3.1/3.2, Vulkan 1.0, OpenCL 1.1/1.2, RenderScript, amongst others.

The RK3399Pro NPU supports 8-bit and 16-bit and is compatible with various AI software frameworks. Existing AI interfaces support OpenVX and TensorFlow Lite/AndroidNN API. AI software tools support the importing, mapping and optimizing of Caffe/TensorFlow model.

Rockchip provides a one-stop AI solution based on the RK3399Pro, including hardware reference design and SDK – speeding AI product development and shortening time to market..

“The age of AI has come. As a global SoC manufacturer in China, we have the market layout of AI for many years. RK3399Pro is Rockchip’s first processor which integrates AI hardware. Its platform can be rapid MP for commercial.” Chen Feng, the Global Vice President of Rockchip, said,

“With super 2.4TOPs performance, low power consumption and abundant interfaces, this product is applicable to various AI application fields such as intelligent drive, image recognition, security monitoring, drones and voice recognition,” commented Chen Feng, the Global Vice President of Rockchip.

If you enjoyed this article, you will like the following ones: don't miss them by subscribing to :    eeNews on Google News


Linked Articles