In order to solve the calculation problem of the world’s most intractable, the Lawrence Livermore National Laboratory, Lawrence Livermore National Laboratory) hope to obtain inspiration from the structure of the human brain, so the laboratory program in this month 31 days to a value of $100 million super computer testing. It should be pointed out that this super computer is the most special place is the integration of the 16 dedicated to simulate the work of the human brain processor.
It is reported that the use of this super computer processor chip called “TrueNorth” by the IBM in 2014 for the first time introduced. In fact, IBM, Qualcomm and Microsoft had previously teamed up to design, develop a computer algorithm to improve the processing speed of the chip. And such as “TrueNorth” is more suitable for this type of chip is applied to advanced artificial intelligence technology, including in-depth study. It should be noted that this chip technology is actually a model of the human brain to run a simulation, although it has not yet been tested, but its future prospects are very attractive.
Specifically, the “TrueNorth” chip and the current server or personal computer used in the chip has a very big difference. The chip set up 5 billion 400 million transistors, which constitute a network of 1 million simulated neurons, which are connected by a large number of analog synapses. Moreover, IBM is mainly used in the United States Department of Defense Advanced Research Projects Agency (DARPA) under the funding to develop TrueNorth chip.
Engaged in truenorth research and development work has been 12 years of IBM’s chief researcher de Harmon della – Modha DharmendraModha said “although in some degree, the structure of the brain looks quite simple, but it can solve some complicated problems.” say。
IBM revealed that it is now five to seven years away from the commercial TrueNorth large-scale investment, but the Lawrence Livermore national laboratory test is an important step in this direction taken an important step.
Similar to the human brain, TrueNorth does not need to consume a lot of energy at run time. For example, the average server chip power consumption of more than 150 watts, while the power consumption of TrueNorth chip is only 2.5 watts. However, TrueNorth can not be used to run the web server, and even can not be completed in a Excel table in the work of several columns of data input. This is mainly because the chip is designed is to perform such as image recognition, or is a kind of artificial intelligence application (such as the recent Google (micro blog) parent company alphabet software AlphaGo beat South Korea chess master Li Shishi that era of computing tasks.
“TrueNorth will be very useful for deep learning applications and for a wider range of machine learning applications.” Van Essen Brian, a computer scientist at the Lawrence Livermore National Laboratory, said.
In fact, Lawrence Livermore national laboratory as early as 2014 began to assess the TrueNorth, but this is the first time the chip has the opportunity to start large-scale testing. Van Essen says his team will be part of the super computer operation gradually transition to truenorth tested, like today, there are many PC will graphics processing of transition to the graphics card to deal with the same principle. He hopes that this technology can help solve problems such as interactions subatomicparticle and network security and other complex issues such as sub atomic particles.
On the other hand, the Lawrence Livermore National Laboratory launched the test for IBM is also an important stage to prove the practicality of TrueNorth.
“They do this very well, TrueNorth may be very efficient, but they still need to prove the accuracy of the model.” University of Washington, associate professor of computer science Sezi – Lewis (Luis Ceze) said.
It should be noted that, IBM is not the only one in the field of computer depth learning to set off waves of enterprise. Currently, Qualcomm is developing its own intelligent cognitive platform Zeroth, the technology allows mobile devices have the innate ability to learn, and do not need to get data through the cloud. For example, Zeroth can identify specific parameters in a picture, so people can use the phone for a specific photography. In addition, users can also use Zeroth technology to identify the picture in the face. More important is that in contact with enough pictures and human face, Zeroth technology will become more and more intelligent.
At the same time, Microsoft researchers are also being launched to allow the programming chip with the company’s Bing search engine to start a collaborative work.
“The race to build a new era of computing chips has already begun. To be honest, we’re still in the lead.” Modeha said.
Finding electronic components, greenfieldcomp.com is a specialized obsolete and hard to find electronic components of the site, in the above can be found more than 30 million pieces of components types, such as: F412R-24MB and psa4531-r, want to know more, please visit our website http://www.greenfieldcomp.com.
