 |
| Intel promises to override Moore's Law within the next decade |
Pat Gelsinger, CEO of Intel, has vowed that the semiconductor industry will overcome Moore's Law (the basic rule that computing power doubles every two years) within the next decade.
This commitment is important because many in the industry believe Moore's Law is no longer applicable and is geared toward improving efficiency rather than hardware.
After 12 years of transitioning from petascal to exascale computing, Gelsinger is challenging his team to reach the next scale, the zeta scale, within five years.
Gelsinger made the above statements in an interview with the founder of Intel and inventor of Moore's Law, Gordon Moore.
According to Moore's Law, the number of transistors on a chip approximately doubles every two years, while the price of the chip remains the same.
This is a bold statement from a company that has worked hard to develop chip manufacturing over the past five years, losing its leadership position to the other two biggest chipmakers, TSMC and Samsung.
It's also a little weird, as Moore himself has said that the law will expire in the next few years as we get closer to corn-chopper ingredients.
This data contains indications that Intel is once again paying attention to developers. The Intel CEO has offered to meet the most extreme needs of ultra-wideband cloud service providers with the promise of advanced features for everyone.
The announcement included discussions about Intel's work with Google to develop the Mount Evans smart processor. and development kits accompanying infrastructure programmers to make network and data center infrastructure programmable.
One of the founders of Intel proposed Moore's Law in 1965
Although it looks like a feature developed by the cloud provider for the cloud provider to use. But high-end organizations such as financial services companies use programmable infrastructure technologies such as the Intel Tofino 3 and P4 network programming languages to achieve high performance in applications such as transactions.
These technologies enable companies to modernize the way they manage their networks and data centers. Now you can program any part of your network and server infrastructure, all you have to do is write a small program. You can choose to do it all the time or only when you need it because it's a program.
These programs are quick to run, which means there is no performance degradation. A few years ago this was out of the question.
Intel is making it easier for more people to take advantage of advanced computing features such as machine learning. This is done through initiatives such as the OneAPI Toolkit and the Intel OpenVINO AI Processor Inference Engine.
With Sapphire Rapids, Intel wants to increase the availability of AI nationwide. It's the codename for a next-generation Xeon processor and promises 30x performance.
This means that AI developers can use general purpose processors. Power accelerators are not very expensive on the market today.