Nvidia's AI chips are advancing faster than Moore's Law, the semiconductor industry's historic benchmark, according to CEO Jensen Huang.
The theory that the number of transistors in a dense integrated circuit doubles every two years is called "Moore's Law", and it was formulated on 19-04-1965.
The “Law” is named after Intel microprocessor co-founder Gordon Moore, who in 1965 outlined the reasons why the number of transistors on a dense integrated circuit would double every year for at least a decade. Looking back at data over the next decade, he revised his “prediction” to put the time it would take for a dense integrated circuit to double its transistors at two years.
The prediction was actually verified, since the number of transistors in an integrated circuit has doubled every two years since then. Moore's prediction, after its practical verification, was called "Moore's Law".
Let's go to today….
"Our systems are moving much faster than Moore's Law," Huang told TechCrunch. Nvidia's chips have improved thousands of times over the past decade, surpassing Moore's Law's prediction of doubling transistor density every year, Huang said.
And he adds:
We can build the architecture, chip, system, libraries and algorithms at the same time. If you do that, then you can move much faster than Moore's Law, because you can innovate across the entire stack.
Moore's Law was so important in the history of computers because it reduced the cost of computers every year.
George is still wondering what he is doing here….
