back to article Europe's scheme to build exascale capability on homegrown hardware is ludicrous fantasy

The EU has declared its intention to build exascale computing capability "on mainly European hardware". This is stated in a European Investment Bank (EIB) report, Financing the future of supercomputing. The 149-page screed investigates the HPC market and supplier ecosystem in Europe with the aim of finding out how it could …

I beg to differ

First of all: I agree 100% with this being a laughable idea in the proposed timeframe.

But: I do think it is time for Europe (including Britain!) to change from a pure consumer perspective of IT technology to a more active and innovative approach.

Am I the only one to hear the crackling in the structure of our beloved and well-understood workhorse architectures? Wouldn't right now be a good point in time to explore alternative ideas while Intel cleans up its mess and the overwhelming part of the big name's budgets go to IP lawyers rather than R&D boffins?

Not all exploration leads to immediate success - but why let that stop us from trying? Let's not pass on a chance to become a more confident player in IT research, maybe make the tech a bit better all over the place and get a speck of "independent of others" (which I tend to think of as a big illusion) to show as well.

21
0

Re: I beg to differ

@LeoP

Indeed!

How long did it take China to go from a standing start?

3
0
Silver badge

Re: I beg to differ

well, ARM is European, though likely soon not EU, so it's not like we would be starting from zero

and while the foundries in Dresden are of GlobalFoundries, so US coroporation, it does show that they can be competitive

so definitely not an insane idea

3
0
Bronze badge

Chris, clearly you have been at ISC and the article starts well.

But think of this like an Apollo Moonshot program. When Kennedy announced that the US would land men on the moon before the end of the decade.. well NASA did not have Saturn 5 rockets sitting in a big shed somewhere....

Europe should be pushing the envelope, and constructing an exascale system. We already have outstanding centres in Germany, in Barcelona, and in Switzerland.

The UK of course has EPCC, Daresbury and RAL labs, Met Office

(I would say ECMWF but they're off to Italy) Plus the ARM based Isambard project in Bristol.

I do get your points about the lack of CPU foundries and DRAM foundries in Europe.

but the chips themselves don't have to be made in European territory.

Remember the titanium for the SR-71 was purchased from the Russkies!

I fully think that Europe can bring new ideas to Exascale. Both in hardware and importantly in the software.

And Exascale is power hungry. Maybe we should be looking at locating this thing somewhere near a green energy resource too.

9
3

This post has been deleted by its author

Europeans are dreaming

Look at Germany's pathetic defense spending and talk of creating their 6 th generation stealth combat aircrafts. American had spent hundred of billion of $ if not thousands to create their F35 ( 5th Gen) aircraft and with their technological know how. I wonder how the Europeans find first the money and secondly technologies to build their futuristic planes. Look at the A400M joint European military transport plane how it ended up into serious cost overran and delay with technical glitches. I really think if Europeans are living in Utopia.

3
18
Silver badge

Re: Europeans are dreaming

Let me fix it for you: American had misspent hundred of billion of $$$ if not thousands to create their F35

Look at Germany's pathetic defense spending

Sure, I have looked. Most of it looks like delivering 1.5-2 times more bang for the buck that UK or USA spending. One of the frigate projects looks a bit in oddball territory (dedicated purely for shipping protection/operations against stone age opponents), but otherwise its procurement can only be admired. It is trying to be functional instead of dick waving.

As far as dick waving, it is not how big it is, it is how you use it. A good example would be the fact that Norway coastal patrol boat squadron can probably deliver more anti-ship capability than the whole of UK fleet.

14
1
Silver badge

Re: Europeans are dreaming

"As far as dick waving, it is not how big it is, it is how you use it."

For dick-waving purposes it's usually how much you spend. Measuring inputs always gives much more impressive numbers than outputs.

8
0
Silver badge

Re: Europeans are dreaming

A recent report showed that German defence procurement is worse than Britains. That is seriously impressive.

6
0
Silver badge

Re: Europeans are dreaming

American had spent hundred of billion of $$$ if not thousands to create their F35 ( 5th Gen) aircraft and with their technological know how.

The thing is, nobody really needs the gold-plated flying radar-evading pig.

What even IS a "5-th generation aircraft"?

The F-35 is tool, the goal of which is to shift taxpayer money to vested interests - and tick all the boxes to make war-fetish think tankers wet themselves on TV. So you get something that is high-maintenance and does not fit any particular realistic mission profile. You can then put a "It's 5th Generation!" sticker on it as if your were in Microsoft Marketing.

You just need a good plane. If need be you can buy Ruskie.

10
6

All well and good, except for the part about Europe being a country like the US.

3
1
Silver badge

Re: Europeans are dreaming

The thing is, nobody really needs the gold-plated flying radar-evading pig.

You forgot to add: "Dead on Arrival on a battlefield where radar no longer works".

If you look at what Russians have done to all of their old aircraft (and are now offering to all existing owners of) as a part of the avionics upgrade is retrofitting some ridiculously overpowered modern ECM. AFAIK Rafale carries a similar package by design.

Jamming on that scale changes quite a few tactical variables:

1. Stealthiness stops being an advantage. In fact, the aerodynamic f*ckups needed to achieve it put all stealth aircraft (and especially ones without vectored power like the F35) at a distinct disadvantage against Gen-4 and 4+ aircraft.

2. Dogfighting ability and training to fight within line of sight (instead of relying on long range missiles) becomes a primary deciding factor.

So in fact I stay by my opinion - the money on F35 was misspent. When stealth came out originally quite a few analysts pointed out that it will be a short lived advantage until jamming is upgraded and installed on all aircraft. We are nearly there now so it is a misspent, not spent money.

3
2
Silver badge

Re: Europeans are dreaming

A recent report showed that German defence procurement is worse

Yes and no.

It has similar or worse overtimes and overbudgets same as all of their procurement. Civil engineering projects, etc. So on that account the report is correct.

The more interesting part is what is being procured. There the comparison is not in Britain's favor. Not sure if that can be extended into the civilian sphere as the Berlin airport and the Hamburg Opera are as much of a white elephant as HS2, but that is procurement anywhere in the world for you. In fact, there at the moment, both Germans and UK trail behind Russia. If we leave all the political elements of what they have done out of the equation and look purely at the procurement and engineering side of the Crimea bridge versus let's say the same airport and HS2 the comparison is clearly not in our favor.

3
1
Silver badge
Thumb Up

Re: Europeans are dreaming

In addition to which you can measure it ahead of time and the inevitable misses and semifailures along the way can be portrayed, for a while at least, as improvements.

1
1
Silver badge

there is no domestic processor or accelerator manufacturer of global scale. No local DRAM players. No European makers of disk or solid-state drives of significance. The same is true for interconnect technology.

True, and this initiative could be an opportunity to change that, if funding is adequate.

In a previous job/life, I was contacted by a governmental agency proposing to enable us to use their HPC means. The offer was interesting but lacked one layer to be really interesting: the software layer. Having hardware is good but without software it's useful only as a heat dispenser.

EIB should also take this point in consideration and promote also adequate software development.

7
0
Anonymous Coward

Remind me again why...

The EU needs to fund HPC for commercial use, if commerce needs it, commerce can pay for it.

Please don't give me the "...but if we don't have what they have we'll be left behind" BS

12
3
Bronze badge

Re: Remind me again why...

AC, you have a point. However, in order to 'push the envelope' and create a new class of machines you do need centrally funded research. You then find the advancements trickle down to the systems which industry uses and pays for.

6
1
Anonymous Coward

Re: Remind me again why...

because, now including the US, much of the world outside of the EU is proving that trade is becoming unreliable and that unreliability could easily including computing hardware. One might also worry that at some point three letter agencies might persuade industries to introduce odd behaviour into chipsets. It could be suggested that some hardware may already contain odd behaviours on purpose.

7
2
Silver badge

Re: Remind me again why...

"Please don't give me the...but if we don't have what they have we'll be left behind" BS".

I am not that convinced I can see that BS all that clearly, I have felt left behind at numerous occasions when I haven't had what "they" have. I won't give you a list of "that" as that might be a bit embarrassing and the list rather long and for many rather familiar, perhaps.

Good for you AC, if you have never experienced anything similar.

As for commerce, a bit of the old "egg and chicken" thing. The EU is a 20 trillion commerce today, who is the chicken who is the egg.

India probably didn't have much of a space program some decades ago, so, is it commerce behind the reality today or is there something else too.

It's Friday after all.

PS. Brits tend to be proud about some world class universities, why, chicken or egg.

PPS. Some virus or chicken just ate my icons.

2
0
Anonymous Coward

Technological advancement...

.. through throwing the towel in and giving up.

Works every time.

12
0
Bronze badge

No European makers of disk or solid-state drives of significance. The same is true for interconnect technology.

Welllllll.. Israel is in the European Song Contest.

So I think Mellanox gets in there too...

12
0
Silver badge

If you mean Eurovision, well, they've let Australia compete, you can't get much further away from Europe than that (Antipodeas of Czechia is just SE of New Zealand), so I wouldn't call being in Eurovision a good precedence for being counted as European ;)

2
0
aks

It's an EU, not European document.

Not only is Israel not in the list, neither is the UK in the timescale being discussed.

I assume it's the French who want to drive this forward using EU money.

1
0
Silver badge

> I assume it's the French who want to drive this forward using EU money.

I'm quite sure that they (in pure monetary terms) contribute more to the EU budget than they get out, so it's more like "using their money with the help of other countries" than the parasitic-sounding "using EU money"...

4
0
Bronze badge

> If you mean Eurovision, well, they've let Australia compete, <

But, as in other areas, they are increasingly xenophobic, and there have been strong statements that Aus should not have been permitted to enter this time around.

1
0
Silver badge
WTF?

What Could Have Been?

Once upon a time such things were being worked on then we had such things as the everyone together ICL a government fabrication. For a while Rodim made disk drives and chips were designed and I believe made even in the UK, certainly transistors were produced here. So we have something of a track record of both chasing the wrong mirages and not doing well in business either.

However, we do need to better harness brains to achieve results to take us forward, remember one of the early computer developments was the result of a tea shop, the LEO computer was a commercially lead development that 'somehow' failed to develop further, neither did the tea shop so we ended up with Starbucks and its elk.

There are other highly desirable developments for which clear commercial needs exist to solve currently emerging problems. Rather than forcing commercial groupings the strategies should push investment into the research needed to answer needs,. Exascale might make nice vanity projects but things to provide food and jobs might be just as useful and could help both fund and support the exotic development. Otherwise the only structures we will deliver will be walls, floors roofs and power, if we are lucky.

4
0
Anonymous Coward

Whatever ever happened to....

The Transputer??

5
0
Bronze badge

Re: Whatever ever happened to....

The interconnect for the Transputer ended up being the Quadrics high speed interconnect.

Have worked with some of those guys, whi are still working in HPC.

11
0

> So EU, find the cash, somehow, give it to Atos

Hahahahaha.

(An ex-ATOS employee)

8
0
Bronze badge

The EU is a dream

Far more easy to do it with Dutch hardware... A joint venture of NXP Semiconductors, Tulip, Philips-Electrologica, KPN Research and TNO could succeed given three decades and TeraBucks in subsidies... make tiny nation rule the waves again.

We would of course be up against the likes of Siemens-Nixdorf, Bosch, ICL, Inmos, ARM, Acorn, Bull, Olivetti, Norsk Data and Nokia and fail to sell outside the Low Countries. Soon South America will overtake Europe.

1
4
Silver badge

RISC-V?

Maybe someone in Europe could design a good RISC-V processor, and send the plans off to TSMC to build? That would give them some European hardware to use. So it's not totally impossible.

4
0
Silver badge

Re: RISC-V?

Maybe someone in Europe could design a good RISC-V processor, and send the plans off to TSMC to build? That would give them some European hardware to use. So it's not totally impossible.

That or one of the other open ISA's like Power, SPARC, MMIX, OpenRISC.

Power and SPARC at least have been commercially used, the others, including RISC-V, don't have much of a commercial history - which could be a good thing.

1
0
Gold badge
Unhappy

Any such plan needs to start by recognizing certain painful truths.

"Exascale computing" is a data centre task

And in the data centre the x86 architecture rules.

Yes it's a fu**ing dinosaur of a code museum which Intel have bolted stuff on over the years to make it look like a real processor architecture.

Consider they tried to clean sheet the problem with Itanium and look where that went.

So you need to start by developing the tools to reverse engineer any ISA and reproduce it with a hard wired design. A modern x86 variant has probably close to a 1000 different instructions.

Eliminating microcode is about the only way you'll get a processor which runs faster than Intel on foundry processes. IOW 6502, not Z80. People will point to ARM but the instruction set for ARM was designed to be easy to map into the control signals to internal PLA's, with few instruction formats. It's the smart way to go, if you have a clean sheet to work with.

Compare that to the multiple, baroque instruction formats of the Intel Pentium whatever of today. You're virtually looking at a context free grammar to decide what one to use right there.

So developing the tools that can implement such an arbitrary ISA would see Europe well placed to design any future ISA they wanted,

Historically the drill has been IS --> set of input patterns --> PALs or microcode -->outputs. But with much greater processing power why not go back to actual hard wired gates, eliminating the gate delays of driving the long lines in the PLA's or the microcode ROM addressing?

IOW a toolset from Arbitrary ISA and register set --> logic design --> logic optimization --> chip layout.

Of course if you can make the resulting chip pin compatible as well that would be better.

I'd start trying to do a single processor without any cache. I think that'll be tough enough.

1
0
Paris Hilton

"Don't bother with Arm:"

So much for my idea of filling a football stadium with Raspberry Pi's, making a huge Raspberry Pi pie.

Paris, coz there's no pie icon, and she's just as delicious.

1
0

I live nearby a firm that makes the machines that make chips. They export all over the world. Not so difficult to set up some of their stuff right here in Europe.

1
0
Bronze badge

Louis, ASML of course. I worked with them last year on their HPC cluster.

HPC being used to design the next generation of machines.

0
0

What about the software?

Great, we can build these machines on some kind of timescale. But without the applications that can exploit them what is the point? Where's the funding for the software development, especially if novel architectures require a large overhaul of existing million line codes?

0
0
Silver badge

Re: What about the software?

Linux is already ported to all the architectures discussed here, and x86 HPC also runs Linux, so that's hardly a problem

the issue is with applications, hand coded assembly in numerical libraries, not with stuff below that

1
0
Anonymous Coward

What about....

...just buying AWS? It's off the shelf, and I'm sure Jeff Bezos would accept a reasonable offer!

0
0

This post has been deleted by its author

POST COMMENT House rules

Not a member of The Register? Create a new account here.

  • Enter your comment

  • Add an icon

Anonymous cowards cannot choose their icon

Forums

Biting the hand that feeds IT © 1998–2018