SC3 card integrates a single BM1682 chip, be competiable with the mainstream x86 server architecture, bringing a new acceleration experience for deep learning.
As Bitmain's second TPU for deep learning, BM1682 has much greater performace
than the 1st generation.
BM1880 is SoC ASIC chip for Deep Learning inference acceleration focusing on edge application.BM1880 TPU can provide 1TFLOPs peak performance for 8-bit integer operation. Lots of AI edge solutions can be implemented by BM1880.
With high performance and low power consumption, SA3 provides more extensive video intelligent applications for users in security and other industries.
With higher security level on the market, providing multiple system-level biometrics including 3D face recognition, voiceprint and speech recognition.
Based on BM1682, SE3 provides cost-effective hardware devices and face recognition algorithms, it is applied to face recognition access, face recognition turnstile and camera.
Utilized for security surveillance, data center, super computer, and robot.
Much higher performance
Word's first buyable AI ASIC chip
Ultra low power architecture，high performance
High performance, low power, provide powerful computing capability for AI industries
High density servers, provide super powerful computing capability
Cost-effective hardware and face recognition algorithm
One BM1682 integrated, New experience for deep learning accelerating
Powerful deep learning for all types of end applications
3D Face recognition algorithm and multiple biometrics system.