Very few of us understand moore’s law or Huang’s Law and its effect in the modern day chip manufacturing in the industry. We will try to explain the laws in this piece of content and as well as try to understand why Nvidia is relying on that law to spread out their research & development as well as future market strategy in the years to come.
Nvidia and ARM Electronics
It was not long ago when we informed you about Nvidia’s latest interest in buying ARM Electronics, the biggest chips and the computers technology manufacturer in United Kingdom. But so far it was assumed that they will be making new processors to compete with AMD and Intel Corp, the two biggest tech giants in the trillion dollar commercial processor market.
Well, that also may be the case but the main reason lies deep, way deeper than what you think as a marketing strategy. We will try to explain in the next part of the content why people are thinking that based on some decade old laws and what might be the implications in computing power.
What is Moore’s Law
It was 6 decade ago, when one of the Intel’s founder, Gordon Moore predicted a ‘trend’ which he think will be observed in the years to come. The prediction was true to every word he said, and so far the whole processor technology is being made following that law, but the scenario may change very soon, read on why.
Gordon Moore said that the densities of transistors on a Chip will be doubled every two years for the years to come. And so far, the speech has proven its point, may be it has been doubled every 18 months now a days, but that don’t make his observation moot.
Moore’s Law in Practice
Now we will take an example here to explain things better. In 2017, Snapdragon 835 was released by TSMC, who are the biggest independent foundry in the world. That chip had density of 52 Millions of Transistors stuffed into each mm square of space, the technology was 10 nm chip.
But this year, Snapdragon 865 was launched in 7nm technology, with over 100 Millions of Transistors per square mmm density, which is almost double that what is was 2 years ago. Thus Moore’s law is satisfied still now, but it will be invalid soon enough. Let me explain why.
Moore’s Law Current Situation
The process or lithography these chips are following now a days is 5 nm to 7 nm process. 7 nm is being used by latest android chipset Snapdragon 865+ and 5 nm lithography signature is of the new Apple iPad Air, which was present latest A14 Bionic Chipset. But the process or lithography can be improved as low as 2 nm chips.
So the way things are looking is that the whole chip manufacturing technology, which solely reply on moore’s law is going to be obsolete in next 2-3 years. Now the alternative, Huang’s law is what Nvidia is following since few years back.
Introduction of Huang’s law
Huang’s law was described by Jensen Huang, Co-Founder of Nvidia. It states that the silicon chips that are used to power Artificial Intelligence, will double their performance every two years. This increase of performance will depend hardware of the chip as well as the improved software used to power Artificial Intelligence algorithm.
Advantages of Huang’s law
The advantage of Nvidia is that they are making only GPU chips for now, which are not required to process a single resource heavy task, but rather multiple simultaneous tasks that are not resource heavy. So with the purchase of ARM Electronics, they can really push down their Huang’s law on all their chip manufacturing technology.
Be with us and subscribe to know what is happening next on this technology, brought to you first by us. And if you have any questions or query, write that down in comment here, we will surely get back to you.
Footnotes : IEEE Spectrum, Move Over, Moore’s Law: Make Way for Huang’s Law, By Tekla S. Perry, November 2012