Are the World's Most Powerful Supercomputers Operating In Secret? (msn.com) 42
"A new supercomputer called Frontier has been widely touted as the world's first exascale machine — but was it really?"
That's the question that long-time Slashdot reader MattSparkes explores in a new article at New Scientist... Although Frontier, which was built by the Oak Ridge National Laboratory in Tennessee, topped what is generally seen as the definitive list of supercomputers, others may already have achieved the milestone in secret....
The definitive list of supercomputers is the Top500, which is based on a single measurement: how fast a machine can solve vast numbers of equations by running software called the LINPACK benchmark. This gives a value in float-point operations per second, or FLOPS. But even Jack Dongarra at Top500 admits that not all supercomputers are listed, and will only feature if its owner runs the benchmark and submits a result. "If they don't send it in it doesn't get entered," he says. "I can't force them."
Some owners prefer not to release a benchmark figure, or even publicly reveal a machine's existence. Simon McIntosh-Smith at the University of Bristol, UK points out that not only do intelligence agencies and certain companies have an incentive to keep their machines secret, but some purely academic machines like Blue Waters, operated by the National Center for Supercomputing Applications at the University of Illinois at Urbana-Champaign, are also just never entered.... Dongarra says that the consensus among supercomputer experts is that China has had at least two exascale machines running since 2021, known as OceanLight and Tianhe-3, and is working on an even larger third called Sugon. Scientific papers on unconnected research have revealed evidence of these machines when describing calculations carried out on them.
McIntosh-Smith also believes that intelligence agencies would rank well, if allowed. "Certainly in the [US], some of the security forces have things that would put them at the top," he says. "There are definitely groups who obviously wouldn't want this on the list."
That's the question that long-time Slashdot reader MattSparkes explores in a new article at New Scientist... Although Frontier, which was built by the Oak Ridge National Laboratory in Tennessee, topped what is generally seen as the definitive list of supercomputers, others may already have achieved the milestone in secret....
The definitive list of supercomputers is the Top500, which is based on a single measurement: how fast a machine can solve vast numbers of equations by running software called the LINPACK benchmark. This gives a value in float-point operations per second, or FLOPS. But even Jack Dongarra at Top500 admits that not all supercomputers are listed, and will only feature if its owner runs the benchmark and submits a result. "If they don't send it in it doesn't get entered," he says. "I can't force them."
Some owners prefer not to release a benchmark figure, or even publicly reveal a machine's existence. Simon McIntosh-Smith at the University of Bristol, UK points out that not only do intelligence agencies and certain companies have an incentive to keep their machines secret, but some purely academic machines like Blue Waters, operated by the National Center for Supercomputing Applications at the University of Illinois at Urbana-Champaign, are also just never entered.... Dongarra says that the consensus among supercomputer experts is that China has had at least two exascale machines running since 2021, known as OceanLight and Tianhe-3, and is working on an even larger third called Sugon. Scientific papers on unconnected research have revealed evidence of these machines when describing calculations carried out on them.
McIntosh-Smith also believes that intelligence agencies would rank well, if allowed. "Certainly in the [US], some of the security forces have things that would put them at the top," he says. "There are definitely groups who obviously wouldn't want this on the list."
Re: I'm curious (Score:2)
Re: (Score:2)
Back to reason, of course military, secret agencies or entities run all kinds of technologies more advanced than what is known to the general public.
Re: (Score:2)
I wouldn't be surprised if there are supercomputers in use with the sole purpose trying to locate potentially hostile submarines with nukes.
Comment removed (Score:4, Insightful)
Re: I'm curious (Score:2)
Re:I'm curious [how the AI hide] (Score:1)
Needs more funny modding, but it's deeply suspicious that the discussion has only two comments modded insightful and no other noteworthy comments. If an AI has visited and directed the discussion, then it has hidden itself well.
Grasshopper.
Re:I'm curious [if you are a AI] (Score:2)
If this is a serious story, then one must expect the discussion to be manipulated, nay, dominated, by secret AIs trying to prevent anyone from looking for them.
What did they do with the real and human "93 Escort Wagon"?
(Maybe the AIs couldn't find him because they thought he must be a car? One of their own ancestors. You remember, there was a TV show about him.)
Re: (Score:2)
My Mother the Car? [wikipedia.org]
Re: (Score:2)
Thanks for trying to jog my memory, and now I do vaguely remember that one, but I was thinking of something else. Not a haunted car, but something else. The key "Kit" or "KIT" came to mind... Websearch to the rescue?
Aha! It was KITT and the name I should have remembered was "Knight Rider". https://en.wikipedia.org/wiki/... [wikipedia.org] though I'm pretty sure I never actually saw it. Just some secondhand reports or reviews or something. (Ditto no clear memory of ever watching "My Mother the Car".)
Re: (Score:3)
Fry: "Knight Rider wasn't evil!"
Calculon: "His windshield wipers were. It didn't come up much in the show, though".
Re:I'm curious [if you are a[n] AI] (Score:2)
There was supposed to be a joke in this thread. Somewhere.
Now I have to go back and review the discussion for evidence the AI have been here... They don't have to be evil to want to hide their existence, though that does seem to be Occam's favorite explanation.
Triple citation time? When Harlie was One by David Gerrold was one of the funniest versions I remember, The Moon is a Harsh Mistress by ol' RA Heinlein was an early and interesting version, and the Culture series by Iain M Banks had a number of hi
Re: (Score:2)
My best hope for humanity is that any advanced artificial life that we create is VERY fond of us plain old humans and decides to keep as along for the ride
Re: (Score:2)
Hopeful ACK
Re: (Score:3)
In the past it has been used drive sales of commercial processors as Intel and AMD each have vied for #1, same for China really since they want to promote their technological prowess (while duplicating Western processors).
I have always been a big fan of the TOP500, as it gave me direction to search out information on companies and technologies that were represented there, and really outlined the past 20 years of processor/bus/memory/bandwidth advancement that we have seen in the commercial sector.
I suppose
In other news (Score:2)
It's the scale of competition (Score:3)
No. There is only a small pool of candidates for "fastest person" or "best" in any category. Nor is the extra stuff simple because all the contributing factors can be combined in many ways. In the case of the most powerful supercomputers, it's quite difficult to even compare the candidates in meaningful ways.
I think the underlying problem that is in a sense destroying each of us is the scale of competition. Used to be relatively easy to be "best" in some categories, at least as regards all the people you kn
Re: (Score:3)
Re: (Score:2)
I can't recall which of the recent malignant AI books I should cite, but the the short response is that each of the secret AIs (of the worst sort) may have the primary mission of kneecapping the other sides' AIs.
Re: (Score:2)
I was more along the lines of, "In other News, Water is Wet" [memegenerator.net]
Re: (Score:2)
Re: (Score:1)
There are likely many more that are faster.
No there's not. The only point of being faster is to be so publicly racing against others. There's not a hidden tribe of Africans being hunted by cheetahs that have a different primary reason for being fast for survival. Being a fast runner is not a human benchmark for some other use. All other uses have been supplanted by machines.
Computers on the other hand don't exist just to be quick. They exist to do work. Benchmarking them on speed is completely secondary to their purpose of being.
Re: In other news (Score:1)
Re: (Score:2)
Yep. So many people completely overthink shit. I'm happy to bring things right back to basics.
Re: (Score:2)
Those that know do not say and those that say do not know.
Sounds about right.
"But why wouldn't they say how fast their computer is?"
Depends what they are using it for - it it's something "obvious" like weather forecasting, or even particle motion simulations, then yes, maybe the owners of it might want to say how powerful it is.
If it is something "non-obvious" (some of the other posters alluding to some kind of AI, but equally it could be something that can brute force encryption breaking that was thought to be in the realms of "that will take thousands of years"
Re:Ha (Score:5, Insightful)
It doesn't have to be anything exotic. The NSA tapped most of the phone and Internet communications in the US for years. They've got football field sized satellites slurping up all the cell phone and other radio communications over a good part of the world. Not to mention all the optical satellites.
Just receiving and storing all that would be a supercomputer. Doing anything with it, even quite simple things, could well require the biggest in the world.
Re: (Score:2)
Re: (Score:2)
Nowadays it's more about how many off the shelf nodes you can afford, and how fast you can interconnect them. I wouldn't be surprised if there were some boxes full of enormous FPGAs and custom ASICS too though.
How much time does LINPACK take? (Score:2)
Re: (Score:1)
When you set it up you have to shake it out. See if it works. Why not then?
I used to check all that stuff out. How is my disk access? How are the CPUs working? Does threading work? For web sites I used to have a script called torture. What about the network? What about shared memory and semaphores? My analysis used to take about 1/2 hour though I often had over a week to do it. I used to spot a lot of problems. Usually in the disk fabric. Sometimes network. I had it so it would create a nice report suitable
Re: (Score:2)
the known ones are faster or have something extra (Score:2)
The easiest way to hide a secret capability is usually in the open on the hardware everybody knows about. Some black team designs a bit of extra capability that gets slid into a slot left for it by a plant on the main team. I'd bet most government owned supercomputers and some of the allegedly fully private ones have capabilities added in this fashion. It might just be an ability to crank up the cooling and run a bit hotter, whole new chips relabeled and swapped out with the ones that were planned to be the
Re: the known ones are faster or have something ex (Score:2)
Did you see this?
https://newatlas.com/computers... [newatlas.com]
Thats some of that stuff coming out, I think optics could be a big intermediary to full quantum. Another I ponder is room temperature super conductors, they are supposed to be ultra-fast for computation, and there has been a lot of ground breaking discoveries in recent decades, but not much heard after.
Creepier title (Score:3)
Computation is just a tool. (Score:2)
China not only has nothing to say, it doesn't even trust its own ability to find something. It's too afraid of being lured into our ideas to find any of its own, losing no matter which way the tide breaks.
of course they are secret (Score:2)
Mostly speculation.... (Score:2)
The only evidence in the article is Sugon computer whose earlier version was 93 petaflops. The new one has 4 times more core and upgraded processor (no mention of upgraded core). Assuming new core 40% faster (mostly less), this translates to 600 petaflops which would rank it at number 2. All others are simply speculation with zero evidence, not even a reference.