The number of data that gets processed today is different from the information fed to old computers a long time ago. Today, everything is digitally available, and as the increased digital data has flooded the present computer system, it is inevitable for it to experience hiccups upon processing. That’s because the current computer system cannot handle the abundance of data that needs consistent handling. That is why IBM researchers have been developing the technology of upgraded computer architecture with a different approach so that it can sustain both the previous and present design.
Results show that the new computer architecture outperforms the conventional computer systems of today’s technology. The Artificial Intelligence technology of computer architecture has found its way towards the mainstream market. The system becomes efficient in handling, storing, and processing elaborate data information from different sources. That’s because the design of the new system mimics the various concepts of a human brain, which by the way is still the ultimate data processor up to date.
Going Deep With Its Development
During the 1940s, a computer architecture called “von Neumann” was built and since then, has become the basis of every computer system. On its structure includes the following essential components. There’s a central processor that acts as the brain of the system where it enables arithmetic and logic executions. There’s also a storage and memory unit that is useful for storing contents and data from external sources. There are a couple of I/O devices (input and output) that is used to enter data to the system. These are the generic parts of a computer system.
However, the researchers from IBM suggested creating a much more efficient system that can achieve 200 times faster performance than its conventional counterpart. From those functions, they focused on a brain-like computer with processing and memory components that coexist. To move forward, researchers based their brain-like architecture design from a specific three levels of the human brain concept.
- The first level is to make use of the device memory state dynamics. It is the same as to how a human brain performs tasks where it involves basic up to complicated recollection of information. The processing of the data happens in the memory function itself. In a human intellect, both the processing and memory occurs in the same location.
- In the second level, the concept was drawn out from the brain’s synaptic network structures where the phase changes ideas of its memory components. This concept was used to accelerate the enhancement for deep neural networks. It allows the computer to make an accurate response and have significant numerical integration.
- Finally, the stochastic nature of neurons and synapses of the brain inspire the third level of the system. Through this concept, the researchers were able to assemble a superior computational substrate beneficial for boosting neural networks and allow the system to have digital-based critical thinking. This phenomenon is a characterized analog storage that resembles biological synapses same as the brain where it stores a lot of data in only one Nanoscale device.
According to the IBM research team, the new computer design bounds to leap a better function than the conventional computing system. The researchers made sure that the new architecture is a good translation of the human brain function. Thus, it will potentially manage, handle, store, and process millions of data without any hassle. And even during the conceptual stage, they already predicted that the new architecture would exceed users’ expectations.
The development of computer architecture is still underway, but the design of the latest innovation is way better than the conventional system. It can process without hiccups flooding information from AI systems, giving it the power to work accurately and quickly.