An American Supercomputer Just Shattered the Speed Record. What’s That Mean?

The world of supercomputing can be quite competitive. In this clip from “The Virtual Opportunities Show” on Motley Fool Live, recorded on May 31, Fool.com contributors Demitri Kalogeropoulos and Jose Najarro discuss a New York Times article that stated a U.S. supercomputer named Frontier recently surpassed one in China to become the world’s fastest. 

Demitri Kalogeropoulos: There’s this machine called Frontier is the name of the computer that is now apparently the first one to have broken the speed barrier, which is 1 quintillion operations per second, which is a crazy number and I’m sure doesn’t give you any context, but this won’t help either. But 1 quintillion is 1 billion billion. [LAUGHTER] That’s a lot.

It’s an interesting article. This is a machine that’s 8,000 pounds, 75 containers basically together. This is interesting, I didn’t know about what some of these supercomputers that used to be used for things like cracking codes, encryption, stuff like that, designing weapons. But today they’re used increasingly for things like developing vaccines, mapping out global climate change, running through extensive, crazy complicated simulations, and design simulations, and things like that.

Apparently so there’s a little bit of a mystery in this thing because there’s apparently some competition mainly between the U.S. and China in terms of who’s got the biggest supercomputer power. According to some estimates, there are about 170 these supercomputers in China versus around 125 or so in the U.S.

Then it’s rumored that there might actually be a faster machine in China is what this article talks about, but they haven’t tested it out and it hasn’t been officially submitted, that’s the way you get ranked in these things. I think we’ve recently talked about Meta Platforms (META -4.58%) working within Nvidia (NVDA -5.95%), I think Jose on securing one of those 125 supercomputers for its own work. I know a lot of companies are into that space too, right?

Jose Najarro: Yeah, definitely. I think Meta is working with both AMD (AMD -4.03%) and Nvidia to make these supercomputers. I think, for me, the Fourth Horseman also made a comment about autonomous driving with Apple (AAPL -3.86%), even though you have big competitors and I think this is the big reason why sometimes even just having that first advantage of data isn’t that important too much right now.

Because right now with a lot of things happening in supercomputers like driving simulations, for example, even if you don’t have that first raw data, if you have a lot of money to be able to buy some of these supercomputers, be able to buy some of these driving simulations, you’ll be able to collect as much data as probably some of the people who already have that experience firsthand.

The other thing I read about this article is that now this supercomputer is about roughly two times faster than the one that was previously so it just shows the huge leap of technology improvement in just the past few years or so.

Kalogeropoulos: Yeah, really cool stuff. Just in the past week, I think I haven’t personally looked at this, but there’s apparently AI-generated things online that you can just describe a random scene and it would AI generate a painting basically, it looks super realistic based on whatever you said. I also saw a headline recently about something that can search for your face, for example, if you look at a picture of your own face it can go through all the public data and just find all this stuff.

So, it’s raising some privacy concerns and things like that obviously. But there’s so much data out there and there are machines now that are going to be able to crunch it in crazy granularity that we’ve never seen before, that’s for sure.

Read More