[HN Gopher] Neuromorphic chip dramatically reduces power require...
       ___________________________________________________________________
        
       Neuromorphic chip dramatically reduces power requirements for
       rolling robot
        
       Author : mdp2021
       Score  : 17 points
       Date   : 2022-06-16 16:55 UTC (1 days ago)
        
 (HTM) web link (techxplore.com)
 (TXT) w3m dump (techxplore.com)
        
       | awinter-py wrote:
       | linear regression dramatically reduces power requirements for
       | rolling robot
        
       | mdp2021 wrote:
       | > _A team of researchers at Tsinghua University 's Center for
       | Brain-Inspired Computing Research in Beijing... has developed a
       | neuromorphic chip that can reduce the power consumption of a cat-
       | and-mouse-type rolling robot by approximately half, compared to a
       | conventional NVIDIA chip designed for AI applications_
       | 
       | > _They built a neuromorphic chip called TianjicX and put it in a
       | small rolling robot called Tianjicat_
       | 
       | > _They also found that their neuromorphic chip-based robot had
       | markedly reduced latency -- [around one fifth of] the NVIDIA
       | based system_
       | 
       | Original research article: _Neuromorphic computing chip with
       | spatiotemporal elasticity for multi-intelligent-tasking robots_ -
       | https://www.science.org/doi/10.1126/scirobotics.abk2948
       | 
       | > _TianjicX, [... a 28-nanometer TianjicX neuromorphic chip with
       | event-driven, high parallelism, low latency, and low power ...]
       | can support true concurrent execution of multiple cross-
       | computing-paradigm neural network (NN) models with various
       | coordination manners for robotics. With spatiotemporal
       | elasticity, TianjicX can support adaptive allocation of computing
       | resources and scheduling of execution time for each task. Key to
       | this approach is a high-level model... which bridges the gap
       | between robotic-level requirements and hardware implementations.
       | It abstracts the execution of NN tasks through distribution of
       | static data and streaming of dynamic data to form the basic
       | activity context, adopts time and space slices to achieve elastic
       | resource allocation for each activity, and performs configurable
       | hybrid synchronous-asynchronous grouping_
       | 
       | It seems the ideas are, given that robots must be in general
       | energy efficient, to both implement a system that allows for
       | resource management, and an adequate hardware implementation.
       | 
       | The article text is not available, and no further details are
       | evident for the neuromorphic side.
        
       | inasio wrote:
       | Dramatically reads to me like one or more orders of magnitude,
       | but it seems it reduced power consumption by 50% when compared to
       | an Nvidia solution; in light of the headline it looks
       | underwhelming.
        
         | kleene_op wrote:
         | > Dramatically reads to me like one or more orders of
         | magnitude, but it seems it reduced power consumption by 50%
         | 
         | That's an order of magnitude... in base 2
        
       ___________________________________________________________________
       (page generated 2022-06-17 23:01 UTC)