terascale

K Computer and Exascale Computing: The New Wave

/ esfein

you brain on k computer

The vast distance between the processing power of the human brain and that of super computers is slowly shrinking. Researchers used K computer, a Japanese petascale computer, to simulate the equivalent of a single second of the brain activity. K computer took 40 minutes to accomplish the feat of simulating approximately 1 percent of the brain’s neural network.

With 705,024 processor cores and 1.4 million GB of RAM at its disposal, the K computer took 40 minutes to model the data in a project designed to test the ability of the supercomputer and gauge the limits of brain simulation.

While computing on this scale is extremely impressive, the abilities of supercomputers are still inadequate in comparison to the insane complexity of the human brain. K computer  is the fourth largest super computer in the world and costs about $10 million dollars to operate annually. The Japanese K computer consumes 12.7 megawatts per hour. According to Fujitsu that’s enough energy to power approximately 30,000 homes.

Related Article: Sweden is Running Out of Trash

The new japanese K computer

860 of these cabinets, working at near full capacity for forty minutes, is equal to 1% of your brains capacity in one second. K computer awe http://www.fujitsu.com/global/about/tech/k/whatis/project/

Your brain, on the other hand, needs only three meals a day, weighs only 3.086 lbs (1.4 kgs), and consumes only 20 watts of electricity. That is a third of the energy used to power your 60 watt light bulb! The goal of large scale computing has always been to create a computer with processing power comparable to that of the human brain. In order to achieve this feat, computer engineers will need to think of the universe as something that never had a box. From Brain-Like Chip May Solve Computers Big Problem: Energy by Douglas Fox:
It is impressive that our computers are so accurate—but that accuracy is a house of cards. A single transistor accidentally flipping can crash a computer or shift a decimal point in your bank account. Engineers ensure that the millions of transistors on a chip behave reliably by slamming them with high voltages—essentially, pumping up the difference between a 1 and a 0 so that random variations in voltage are less likely to make one look like the other. That is a big reason why computers are such power hogs.

The Neurogrid computer, developed by Kwabena Boahen of Stanford University, aims much smaller than the K computer and by virtue, much larger. While the traditional computer is strict and rigid, the Neurogrid computer is designed to accommodate for the organic nature of the brain. Instead of utilizing the efficient methods of other computer engineers, Boahen attempts to hone in on the organized chaos of the human brain.

Related Article: Electronic Brain Implants Increase Intelligence

The Neurogrid and other computational innovations such as the K computer are likely to usher in a new wave of high level processing where emphasis on power is transferred to an emphasis on the delicate balance between information, energy, and noise. Perhaps by blending the two schools of thought- Large, impressive and small, noisy– we will be able to power machines capable of processing on a similar plane as the human brain. Incorporating efficient practices will also aid in the removal of moral obstructions. Should we power this mega-rad pc or provide the homes of an entire suburb or small village with electricity? Well with efficiency-not to be confused with that dreaded bureaucracy- we may be able to do both!
The dreams of computer engineers are likely to come to fruition very, very soon.
If petascale computers like the K computer are capable of representing one per cent of the network of a human brain today, then we know that simulating the whole brain at the level of the individual nerve cell and its synapses will be possible with exascale computers – hopefully available within the next decade.
For a better understanding of the difference between mega, giga, peta, exa, etc. just remember that each one is 1000 times more than the last.  So, an exascale computer is 1,000 times more powerful than a petascale computer (like the K computer), 1,000,000 times more powerful than a terascale computer, and 1,000,000,000 (billion) times more powerful than a gigascale computer (the computers that you and I have).

Related Article: The Singularity is Nigh Upon Us: The Merging of Humans with Technology

 

Sources: