News
News
Contact Us
Contact Us
Shenzhen Office:208,H Block,Baoan
Zhigu Industrial Park,No.4 Yintian
Road,Xixiang Street, Baoan District,Shenzhen,Guangdong PRC.
Meizhou Factory:AD6 Building,Dongsheng Industrial Park,Economic Development District,Meizhou,GD,PRC
The development of the United States and human brain circuit board has a computer 9 thousand times the speed of computing
- U.S. scientists developed a new type of circuit board Neurogrid, using 16 custom Neurocore chips, the cost of production reached $40 thousand, the number of neurons and synapses can be simulated far more than the same energy consumption in the case of other brain simulations performed on the tablet.
- The 16 year old Patrick Khanna, lost his left arm for meningitis. In Livingston, the new bionic prosthetic hand touch installation, manufacturing company. The researchers say the Neurocore chip can improve the ability of the disabled to control the prosthetic limb.
- Beijing time 13, according to foreign media reports, the United States scientists have developed a new circuit board, in the design can work as the human brain. The circuit board is called "Neurogrid", which is capable of simulating 1 million neuron cells as well as billions of synaptic connections. Experts say the speed of the circuit board can reach 9000 times that of the current computer. The size of the Neurogrid is comparable to a iPad, making the cost of $40 thousand, and the number of neurons and synapses that can be simulated is much more than other brain simulations performed on a tablet computer with the same energy consumption. Neurogrid's research and development work was funded by the National Institutes of health.
- The researchers said the computer is simply not worth mentioning compared with the human brain. The operation speed of the mouse cerebral cortex can be 9000 times as much as that of the personal computer simulating its function. Associate professor of biological engineering, Stanford University, Bergen - Cova Bernar in the Institute of electrical and Electronic Engineers Association, said in an article pointed out that, in addition to slow computing speed, the energy consumption of personal computers is much higher than the brain, which can reach 40 thousand times the latter. He said: "from the point of view of energy consumption, the computer is difficult to compare with the brain." Bergen's article describes how American and European researchers have made use of silicon and software to simulate the electronic system that can be used to model neurons and synapses.
- The development of Neurogrid by Peter Hearne and his research team, using Neurocore chip 16 custom. These chips are capable of simulating 1 million neuron cells as well as billions of synaptic connections. In the design, Neurocore chip has a very high energy efficiency. Bergen et al., the strategy is to allow the determination of the "synapse" shared hardware circuit. The size of the Neurogrid is comparable to that of a iPad, which can simulate the number of neurons and synapses that are much more than other brain simulations performed on a tablet computer, much more than the same energy consumption. Neurogrid's research and development work was funded by the National Institutes of health.
- In neurogrid prototype development, primary Hearn's team will further study, reduce costs and R & D compiler software, let not have neuroscience knowledge engineers and computer scientists can also use neurogrid to solve the problems, such as the control of humanoid robot. The advantages of Neurogrid in terms of speed and energy consumption make it an ideal tool, far more than simply modeling the human brain.
- Currently, the uncle Hector Enzheng with Stanford University's other scientists cooperation, for paralyzed patients developed the use of prosthetic control similar Neurocore chip. He said: "at present, you have to know how the brain works, for this kind of prosthetic programming. We want to develop a neural compiler that can use this circuit board without knowing the synapses and neurons." At Stanford University's office, the $40 thousand Neurogrid prototype was set on the table. He said the possibility of a significant reduction in costs is the existence of.
- Neurogrid uses 16 Neurocore chips, each block can be simulated 65536 neurons. With the aid of modern manufacturing technology, it is only $400 to make a Neurocore chip for the control of prosthetic limbs. In the article, Bergen pointed out that the energy consumption of Neurogrid is only about 1 million of the personal computer that simulates 1/100000 neuron cells, but it is still far more than our biology CPU. He said: the human brain computing speed is 80 thousand times the Neurogrid, energy consumption is only 1/3 of the latter. A significant reduction in energy consumption at the same time providing a more configurable and scalable energy consumption is the ultimate challenge for the nerve engineer.
- Last one：3 men burn 30 tons of copper circuit board jingfangxingju
- Next article：After 85 Indoorsman make magic board