By submitting this form, you are agreeing to the Terms of Use and Privacy Policy.
Although analogue computing is an older technology than digital computing, it was mostly ignored as digital computing gained popularity.
Numerous new ideas and technologies have emerged as the industry pushes for lower power and greater performance processing for machine learning (ML) and artificial intelligence (AI).
Among them, analogue computing has come back to life as a fascinating strategy for more effective processing. Nevertheless, there is a lot of space for development because the technology is still very new for this particular application.
The narrow dynamic range, different transistor biasing regimes, and temperature changes of analogue chips prevent them from scaling as well.
Better privacy and security for consumer devices are also provided by the information processing on edge AI chips. These chips enable for the local processing of massive volumes of data, which lowers the chance that private or confidential information will be intercepted.
For instance, security cameras with on-device computing capabilities can lower privacy threats by analysing the video locally to identify the pertinent video segments to be sent to the cloud or data centres.
Along with 5G networks and other ongoing advancements, the consumer electronics and automotive industries have both experienced tremendous growth.
Instead, analogue computing enables in-memory computing, which permits data to be processed in the same location that it is stored.
An industry-emerging trend has been the major influence of data transportation energy as data speeds have increased, process nodes have become smaller, and global interconnects have become longer.
The physical transfer of data into and out of memory has evolved into one of the major drivers of chip power consumption because of rising parasitic.
IISc is a leading mobiliser of the equipment in the market. The latest integration has been the Next-generation analogue computation chipsets might be speedier and use less power than the digital chips found in the majority of electronic gadgets thanks to this design architecture.
The group has created a prototype of an analogue chipset named ARYABHAT-1 utilizing their innovative design framework. Considering Artificial Intelligence (AI)-based applications like objects or natural language processing, this kind of chipset can be extremely useful.
Mythic is part of the component manufacture trending companies in the current industry. The M1076 Analog Matrix Processors is a single-chip analogue computing system that can give powerful computing power while using very little power.
With a 3W power profile, each AMP chip can produce up to 35 TOPS (trillions of operations per second). With 10 TOPS per watt, Mythic would be in the driver’s seat among the expanding group of edge AI platforms.
This will primarily concentrate on edge AI installations where big or many models are needed and power and performance are crucial.
