High-precision Computing and AI Board Card

With the constant expansion of high-precision computing and artificial intelligence applications, the need for efficient completion of various complex calculations has emerged in all aspects of life and work.

The high-precision and high-performance computing board and AI board of Lingyange adopt advanced packaging architecture, fully supporting high-performance computing and low-power requirements such as FP32, FP16, INT8, etc., which can be used in various signal processing applications, accelerate AI operations, and improve computing and power efficiency.

High-precision and high-performance computing board supports OpenCL, providing users with CLBLAS, CLFFT, CLRNG, CLSPARSE and other function libraries, as well as operator customization and advanced optimization.

Lingyange’s AI accelerator board provides model adaptation from our model zoo, and has already supported the completion of 10+ performance optimizations (all models in the yolo series have been optimized).

Detailed specifications are as below :