Europe enters the exascale supercomputing league with Jupiter | Hacker News Reader

Europe enters the exascale supercomputing league with Jupiter

ec.europa.eu

71 pointsSami_Lehtinen43 comments

"With JUPITER, Europe becomes the home of the most powerful computer in Europe" -- Henna Virkkunen

What does that mean?

My guess is that it means that not only does "Henna Virkkunen, Executive Vice-President for Tech Sovereignty, Security and Democracy" write her quotes with AI without reading them carefully, but that she's also doing it with an older model.

Maybe they wanted to say “with [JUPITER, Europe],” as in “CERN, Switzerland.” This would still be unnecessary and confusing, but that’s the most sense I can make of it.

I am assuming before JUPITER the most powerful computer in Europe was not at home in Europe?

To quote zefrank: The tallest giraffe is taller than all other giraffes

I took it to mean that the controller of the most powerful computer in Europe is Europe, as opposed to Apple or Google or another private entity.

I think it’s pretty clear if you read the rest of the sentence, which you’ve omitted: “and the fourth most powerful in the world”. If she hadn’t said “in Europe”, the whole top 3 might be in Europe as well.

I mean, it's a straightforward enough statement. But what was happening in Europe before..?

Confusing grammar aside, the most powerful computer before this was HPC6 in Italy which is definitely in Europe.

https://www.top500.org/lists/top500/2025/06/

https://en.wikipedia.org/wiki/Supercomputing_in_Europe

Your Top 500 link says something completely different:

"The 65th edition of the TOP500 showed that the El Capitan system retains the No. 1 position. With El Capitan, Frontier, and Aurora, there are now 3 Exascale systems leading the TOP500. All three are installed at Department of Energy (DOE) laboratories in the United States."

HPC6 is listed below those, and far less powerful.

In Europe, HPC6 was the fastest before Jupiter. It was No.4 last year and 6th now that JUPITER is taken 4th place.

Yes, but that fact does nothing to make the statement in question more sensible, so I don't know why you bothered to point it out, or to emphasize the "HPC6 [is] in Italy which is definitely in Europe" aspect. (Although I don't think the statement in question was all that nonsensical to begin with — IMO, it was just inartfully phrased.)

Sounds like a dril quote.

I believe this refers to the fact that, before Jupiter, the most powerful computer in Europe was HPC6, which is owned by Eni S.p.A, a private company, and the following most powerful computer was the Alps system, located in Switzerland. So, in this context, this new supercomputer is owned by Europe, “the public” and it’s located in the European Union sovereignty within Europe, the continent.

Edit: I found the full quote in the website of the Jülich Development Center and I guess it makes sense why it was editorialized for the eu-wide website.

“This is a historic milestone. With JUPITER, Europe is reaching the highest level of high-performance computing. JUPITER is also a testimony for Germany's long leadership in HPC. Today, it became the home of the most powerful computer in Europe and the fourth most powerful in the world. From European perspective, JUPITER is a pioneer. It shows that when we combine national vision with European cooperation, we can achieve global excellence.”

Source:

https://www.fz-juelich.de/en/news/archive/press-release/2025...

How do the gpu training clusters rank against the traditional “largest supercomputer” metric? Or is that comparing apples to oranges?

I think by some measures XAi already has the largest (200k H100 equivalent) and already constructing the 1 million gpu cluster.

From the add I did over on WP Current Events, based on the June preliminary system test (data from Top 500 https://top500.org/lists/top500/2025/06/):

"It is currently being commissioned and has achieved a preliminary HPL value of 793.4 Petaflop/s on a partial system."

Top 500 Rank: 4 (behind El Capitan, Frontier, and Aurora at the DOE)

System Config: JUPITER Booster - BullSequana XH3000, GH Superchip 72C 3GHz, NVIDIA GH200 Superchip, Quad-Rail NVIDIA InfiniBand NDR200, RedHat Enterprise Linux

Manufacturer: EVIDEN https://eviden.com/solutions/high-performance-computing/hpc-...

Location: EuroHPC/FZJ, Germany https://www.fz-juelich.de/en/ias/jsc/jupiter

Cores: 4,801,344

Rmax (PFlop/s) / Rpeak (PFlop/s): 793.40 / 930.00

Power (kW): 13,088

The actual technical details,

https://www.fz-juelich.de/en/ias/jsc/jupiter/tech

something with a little more technical detail:

https://www.fz-juelich.de/en/ias/jsc/jupiter/tech

looks like its just a big Nvidia installation. IB, Ceph, K8s, H100s

If that’s “so simple”, why we don’t see everyone “just” doing the same?

It might just be expensive... Some other countries might just not need those when they are struggling to get fresh water...

And so on, but that's just my 2 cents.

Europe is not a country.

Federal Europe is a proto-country - and we are already citizens of the European Union. In short: Europe is my country !

"Everyone" doesn't apply just to continents, isn't it?

To be fair, being "just" a big Nvidia installation doesn't really mean that it's "so simple" (person you're replying to didn't claim it was simple), considering that every "Nvidia installation" has a slightly different set of factors like node count, per node memory, CPU choice, and networking configuration. All it says is, like many (most?) other modern supercomputers, it's a giant NVIDIA GPU cluster.

> a 21 Petabyte Flash Module (ExaFLASH) is provided based on the IBM Storage Scale

Nice. How much does this cost?

> The scratch storage is based on 20 IBM Storage Scale 6000 systems utilizing NVMe disk technology, based on the IBM Storage Scale solution. With 29 PB of raw and 21 PB of useable capacity

That's also a lot of waste, isn't it?

With 29 PB of raw and 21 PB of useable capacity.

RAID gonna RAID.

Don't know why your're downvoted, this is valid criticism. EU just bought 500 mil worth of hardware from the US, no specific European innovation, apart from system management (ParTec Italy). But would be happy to be proven otherwise.

Who exactly were you going to buy equipment from? AMD/Nvidia/Intel are the only players in town.

China with SMIC and Huawei's Ascend 910C would be an example of what you do if you want to pursue strategic autonomy. It's like asking "who exactly were you going to buy a 6th generation fighter from?"

Without questioning this idea on political grounds, I am not sure if it would be at all possible for an Ascend 910C cluster to enter the supercomputer rankings. I could not find a public datasheet on this chip (would appreciate a link), but my impression is that it is an AI accelerator that does not target FP64, whereas TOP500 is looking at HPC (FP64) performance [1].

[1] https://top500.org/resources/frequently-asked-questions/

Until a year ago, few thought there would ever be a need for Europe to have strategic autonomy from the US in an area that was already solved by trade agreements. Presumably this project was conceived long before that.

It was probably more like system designed in Europe, made in China, assembled in Europe thing.

Sure, the HQ of the core component(Nvidia) is incorporated in US but that too is a multinational effort.

Germany got European taxpayers to buy them US hardware. That's a fairly normal day in the EU.

US Hardware, with essential components manufactured in Taiwan with European technology.

Don't forget it's made in china.

Seems like it has 5 petaflops of compute from a SiPearl (French) designed Rhea-1

Edit: that might be speculative or a future addition cause they taped out less than 2 months ago https://www.eetimes.eu/sipearl-tapes-out-rhea1-processor-clo...

Which was delayed 2 years. I’m speculating this was supposed to be mostly or exclusively this but they needed a computer now. Or needed to spend the budget now.

So US supercomputers aren’t an achievement because they just bought chips from Taiwan and those aren’t also an achievement because they bought photolithography machines from ASML in Europe?

Since when are supercomputers anything other than just the ability to afford a lot of hardware?

Honest question: what will likely be done with a computer like this?

Find websites that don’t properly handle GDPR

This cracked me up, I'm absolutely dying

crack passwords?

Looks like a research cluster. Universities / their Researcher can apply for reasources instead of setting up own hardware.

Quite a varied field of research. From Computer Science and Neuroscience to Climate Research and more. You can find an overview for some of the HPC systems here: https://www.fz-juelich.de/en/ias/jsc/systems/supercomputers/...

Reply on news.ycombinator.com