Chinese scientists in a recent breakthrough have developed a new analog chip, capable of performing calculations on its physical circuits instead of using binary digital processors.
The most distinguishable feature of this new AI chip is that it can outrun AMD and Nvidia’s cutting-edge GPUs by 1000 times.
In a research study published in the journal Nature Electronics, the researchers also shed light that this chip also solves “century-old problems” of poor precision and impracticality.
The new Chinese analog AI chip has successfully tackled bottlenecks related to energy and data processing that scientists face in the fields of 6G and artificial intelligence.
When it comes to energy efficiency, the chip also matches the precision of standard digital processors while utilizing 100 times less energy.
The device also consists of novel analog design, showing arrays of resistive random-access memory (RRAM) cells.
According to researchers, "With the rise of applications using vast amounts of data, this creates a challenge for digital computers, particularly as traditional device scaling becomes increasingly challenging.”
"Benchmarking shows that our analogue computing approach could offer a 1,000 times higher throughput,” they added.
The chip also achieves efficiency by processing data within its hardware, thereby accomplishing 100 times better energy efficiency.
