UK axes plans for Edinburgh-based exascale computer

The UK's £1.3 billion ($1.66 billion) plan for AI and tech investment that included an £800 million ($1 billion) exascale supercomputer at Edinburgh University has gone up in smoke. The blueprint for Edinburgh's computer was revealed in October 2023, and described as "game changing" by the then administration. The game has …

[Anonymous Coward]

Of Course The UK Has No Need At All For Infrastructure.......

....you know.....HS2......

....you know.....modern computing.....

....must be true......we have bean counters who say that it is so!!!

Counter Example:

(1) The first Forth Road Bridge cost less than £30 million in 1964

(2) Treasury estimates show BILLIONS in revenue generated in Fife since 1964

....and only a few miles from the cancelled computer.....

[Steve Button]

Re: Of Course The UK Has No Need At All For Infrastructure.......

False equivalence.

(and as an aside... HS2? Seriously? I was on the fence about it previously, but since the pandemic passenger numbers are way down. It's just not needed right now. Enough people have realised they don't need to travel as often to get stuff done.)

[Anonymous Coward]

Re: Of Course The UK Has No Need At All For Infrastructure.......

(Different AC)

Funny how the plans for pylons and pipes to get electricity and water into the Home Counties are going ahead. I've no doubt that whenever this exascale project gets resurrected it will end up in London, Oxford or Cambridge. Important People don't want to be traipsing about oop North y'know.

[Like a badger]

Re: Of Course The UK Has No Need At All For Infrastructure.......

"Funny how the plans for pylons and pipes to get electricity and water into the Home Counties are going ahead"

Errmmmm...maybe that's where demand is, and supply is elsewhere, as part of policy decisions already made through the democratic process. And this business of long infrastructure lines has always been the case - Birmingham City Council had to build some very big pipes 73 miles to Wales to secure reliable, safe drinking water, and that was in 1896, Liverpool did something similar. The UK national grid was conceived in 1926, to address the fact that electricity system was fragmented and inefficient, and the then-sources of coal were a very long way from most of Britain's cities.

"I've no doubt that whenever this exascale project gets resurrected...."

We've yet to see any credible evidence of the benefits of exascale computing, and throwing a billion quid at the matter just because everybody else is doing it seem pointless, especially as the public finances are in a very sorry state. By all means let's see it resurrected when somebody's found a use beyond national willy-waving.

"....it will end up in London, Oxford or Cambridge"

Why? Labour want to hang on to all the Scottish votes they recently acquired thanks to the SNP's self-immolation. And Edinburgh is a fairly obvious choice aside from politics - the Uni already have one of the UK's most powerful computers, along with globally recognised strengths in information technology and commercialisation thereof. London should be a non-starter for reasons of cost, energy supply and no need for public bungs, Oxford should be out because Cambridge is better at science, and Cambridge should be out because (along with Oxford) they're all research focused egg heads who have plentiful funding anyway*.

* If anybody is not offended by my generalisations there, then do get in touch and I'll see what I can do. Disclosure: I'm not Scottish, neither affiliation nor animosity to any of the seats of learning mentioned.

[Ian Johnston]

Re: Of Course The UK Has No Need At All For Infrastructure.......

* If anybody is not offended by my generalisations there, then do get in touch and I'll see what I can do. Disclosure: I'm not Scottish, neither affiliation nor animosity to any of the seats of learning mentioned.

I am Scottish with links to all the universities you mention and I endorse your message.

[Korev]

Re: Of Course The UK Has No Need At All For Infrastructure.......

(and as an aside... HS2? Seriously? I was on the fence about it previously, but since the pandemic passenger numbers are way down. It's just not needed right now. Enough people have realised they don't need to travel as often to get stuff done.)

The big problem that HS2 will solve is the lack of rail capacity north of London, sadly this is hard to communicate so they focused more on getting out of London/Birmingham (delete according to prejudice) a bit quicker...

[TechnicalVault]

Re: Of Course The UK Has No Need At All For Infrastructure.......

Short-termist take.

HS2 is not merely about passenger numbers for the next few years, it's building rail for the next 150 years. Most of our existing infrastructure is over 100 years old so we need to address the technical debt in the rail network caused by being the early adopters. By upgrading from a loading gauge of W10 on the West Coast mainline to UIC GC we allow the line to take refrigerated ISO containers and modern sized trains. To change the loading gauge on existing tracks will take rebuilding ancient tunnels and viaducts, plus moving overhead cables and changing the distance between tracks. If you look at existing rail most of it was built in the mid 1850s, we don't often get a chance to do these projects so yeah it's going to be pricey, but it lasts.

[David Hicklin]

Re: Of Course The UK Has No Need At All For Infrastructure.......

>> loading gauge

But what do you do when you get the the line at either end? You are back to normal tracks then.....

[TechnicalVault]

Re: Of Course The UK Has No Need At All For Infrastructure.......

>> But what do you do when you get the the line at either end? You are back to normal tracks then.....

In the interim... If you're a passenger train, you either unload your passengers (lots of people want to go to Birmingham, Manchester and London) or carry on over HS1 to Europe. If you're a goods train (I suspect there won't be as many of those by day) then you put the containers on lorries as is, that's the whole point of containers.

However, this was not supposed to be an isolated project. The idea was to replace the whole of the UK's mainlines with a series of high speed up-spec'd lines just like the French and Chinese did. Like I said, these lines have (Ship of Theseus style) lasted hundreds of years, you need to think of it as a series of long term projects. Whether it'll happen with current attitudes I'm not so sure,

[Bonzo_red]

Time for a change of name

Perhaps a petition should be started to rename DSIT to the Department of Science, Innovation and Silly Technology, or DeSIST.

[collinsl]

Re: Time for a change of name

About as good as when they tried to re-brand the Department of Trade and Industry to the Department for Productivity, Energy and Industry, which would have been shortened in some circles to the Department for Productivity, ENergy, and InduStry. Or just plain "dippy"

[m4r35n357]

Good

Ignore the artificial idiot hype, and do it properly.

[Jellied Eel]

Re: Good

Ignore the artificial idiot hype, and do it properly.

By properly.. I have a modest proposal. Implement an AI Tax whereby every time a company mentions AI, they're taxed say, £100. There may soon be funding for a superdoopercomputer.

Otherwise it seems to become a simple business case. Why does Edinburgh need £1bn in tin, other than to maybe claim a spot in the supercomputer rankings? What could it do that couldn't be done in a rack of CPU or GPUs? Especially when for AI to become useful, it needs to be possible in something that costs considerably less than £1bn.

Which I guess is the usual challenge with research funding. Demonstrate some practical benefits in exchange for public money. ISTR Edinburgh became a supercomputer centre mainly on the back of oil & gas, which Labour's in the process of banning. And I guess that challenge also extends to researchers putting together proposals to try and book time on any supercomputers to conduct their own experiments. At least good'ol RACF made that somewhat easier in the good'ol mainframe days we're recreating with all the cloudybollocks and AI hype.

[HPCJohn]

Re: Good

"What could it do that couldn't be done in a rack of CPU or GPUs" Supercomputers ARE composed of racks of CPUs and GPUs

I guess what sets them apart are the networks which tie those CPU/GPU systems together and the high bandwidth storage.

[werdsmith]

Peter Kyle MP, boss of DSIT with his doctorate in "community development" is a champion of AI and wants to force tech firms to share their test and training data with the government.

At this rate he will force the tech firms offshore.

[Doctor Syntax]

I'm old enough to be reminded of Harold Wilson and his "white heat of technology" looking around for a big, advanced tech project to cancel.

[m4r35n357]

Bye-bye Inmos, nobody needs a transputer!

[Doctor Syntax]

That was later. I suspect its problems were too expensive and just too different for the industry to handle.

Wilson's statement of his true colours was cancellation of TSR2.

[Ian Johnston]

Yeah, because if he hadn't cancelled it we might have won the Cold War. Oh, hang on ...

[m4r35n357]

Fair enough, memory blurry, I was a lot younger then, just looking to go start an engineering degree as pretty much the whole industry (MOD aside) was being cancelled or sold off.

I grew up with talk of concorde, harrier jump jets, blue streak, polaris, fast breeder reactors, microcomputers.

France & Germany managed to keep their engineering industries & prestige projects; we just gave up/sold out (Thatcher).

[druck]

All the big defence projects were cancelled the decade before under Labour, but don't let that stop you blaming every wrong decision in history on Thatcher.

[Anonymous Coward]

everything leading to now comes from thatchers short term profit taking and selling the countries assets.

Tories do this shit every time to fill their own pockets.

[Anonymous Coward]

As opposed to the Labour party, who for example instructed Rolls Royce to ship then state of the art jet engines to Russia back in 1946. According to current RR employees, these were never paid for, and the sale was a government instruction. Nothing like helping your enemies out with free technology.

Those engines went to become not just the foundation of Russia's military engine industry, but China's.

[werdsmith]

Stafford Cripps. In fact Soviet engineers were trained at Moor Lane Derby to support these engines and they developed them into the Klimov in the MIG 15

However, the Nenes were becoming obsolete. The easier to manufacture centrifugal compressors with their large frontal area were not ideal for fast jests and so both sides of the cold war had to developed axial flow gas turbines. RR Derby had the Avon in development and that was the future.

[collinsl]

Yes, but that's stark comfort to the allied pilots shot down over North Korea by Nene-powered MiG jets which Russia sold to China and North Korea.

And those who were then shot down in the early Vietnam war by the same jets.

[m4r35n357]

Thatcher converted us to servile industries by selling off _civilian_ assets, and ignoring engineering altogether - I have never aspired to work in defence myself, I am just saying what was there when I was younger. The labour party has been a huge disappointment over the years, but nothing compared to the sheer malice of the tories and those they bribed with tax cuts.

[Like a badger]

Actually, it wasn't just civilian assets - DERA was privatised and became Qintetiq (however that's spealt), and whilst after Maggie, the Tories continued that with things like privatising the former military run air sea rescue, with privately provided air refueling tankers, privately provided maintenance of everything military, sale of every possible military base for the land, etc etc.

If the Tories have any of their grandmothers left alive, you can be sure they'll have been packaged up as collateral for some tax avoidance vehicle that's then been sold on to greedy but hapless investors.

[Doctor Syntax]

"I was a lot younger then"

We all weere.

[Doctor Syntax]

Does anyone care to list the actual, genuine STEM graduates we've had as PM?

[codejunky]

@Doctor Syntax

"Does anyone care to list the actual, genuine STEM graduates we've had as PM?"

Thatcher springs to mind.

[m4r35n357]

Re: @Doctor Syntax

Yep, it is not a solution. The political process is designed to corrupt and manipulate anyone who takes part, however well-meaning they might be.

Of course, Thatcher was never well-meaning, and managed to bag herself a ludicrously rich hubby, so had no personal stake in the UK economy, or science & technology.

[abend0c4]

The decisions have not gone down well

I'm not sure the extent to which government investment makes a lot of difference.

The last major technology player to emerge around a University was ARM - and I think that had a lot more to do with the individuals than the University itself. There certainly wasn't a lot of investment in the Computer Lab at that time. And ARM grew too big for UK investment to sustain. That, ultimately, is the gotcha.

[Peter Gathercole]

Re: The decisions have not gone down well

What grew out of Cambridge University was Acorn Computers, the inventor of the Acorn RISC Machine.

ARM is an interesting development that out-lived it's parent.

[Bonzo_red]

Re: The decisions have not gone down well

Manchester University was heavily involved in the more modern ARM processors, I believe.

[collinsl]

Re: The decisions have not gone down well

Manchester university was heavily involved in computing full stop, Alan Turing being a lecturer there after the war.

[codejunky]

Hmm

And I am sure there will be lots more cuts to come. As a country we have a bad financial situation and leaders with spending problems (or problems with not spending).

[Like a badger]

Re: Hmm

"As a country we have a bad financial situation and leaders with spending problems (or problems with not spending)."

No, we have a democracy problem - people want public services like Denmark, and taxes like Chad, so that's what politicians promise them, unfortunately it cannot be done. There's nothing inherently wrong with either a big state, nor with a small state, nor even with an occasional large budget deficit, or a chronic but very low deficit. Unfortunately successive British governments have given us chronic large deficits.

The average budget deficit over the past fifty odd years is 3.7% of GDP, it needs to be nearer 1% to minimise the problem of debt taking up ever increasing amounts of government spending. The UK is in exactly that problem place now - that government has to spend £89bn on debt interest, and therefore has to curtail current and capital spending, yet is increasing the debt pile by around £1bn a week as of this year.

Tax take is around 41% of GDP last year, if we take that to 44% of GDP, we would close the deficit to around 1% (I'm mixing some current-ish and average figures hear, bear with the argument). However, that 44% tax take means an overall increase of 7.3% in tax, we already know that's not happening. Or we can both raise taxes and cut spending, but where is there any national consensus as to where we can save (say) half the amount required through cuts - that 50% would be around £50bn reduction in government spending each and every year. There's only so many things the population will accept cutting, although super-computers for academia are one, of course. I'm sure individually we could nominate £50bn of recurring cuts (bear in mind that still requires a permanent 3.6% rise in taxes), but the challenge is getting a national consensus - my £50bn of cuts won't be popular with many other people. And regardless of people, politicians have their own agendas where they turn a tin ear even to their own voters.

[Doctor Syntax]

Re: Hmm

What not to cut - and, if possible, raise spending on - is stuff that stands a chance of increasing GDP in the future. Whether the Edinburgh supercomputer or any of the other cuts would have fallen into this category I have no idea. But one thing I'm sure of - no government is going to be willing to invest in anything likely to start bringing in GDP increases in the next electoral cycle rather than the current one.

[codejunky]

Re: Hmm

@Doctor Syntax

"What not to cut - and, if possible, raise spending on - is stuff that stands a chance of increasing GDP in the future"

Unfortunately this tried and tested method has led us to our current financial situation. We have lots of spending, high taxation and calls for more spending to boost future growth. It sounds increasingly like a gambling addict asking to borrow more because they are certain they will win next time.

The gov can spend more to make GDP figures look good but it is just a façade.

[collinsl]

Re: Hmm

I'd like to see your list of stuff to cut, considering that cutting has been happening since 2010 and there's not much left which can be cut now without essential public services being pushed beyond the brink of actually performing their roles at all.

[codejunky]

Re: Hmm

@collinsl

"considering that cutting has been happening since 2010 and there's not much left which can be cut"

This doesnt really make much sense with the current situation. You say cutting has been happening but spending only reduced slightly 1 year. That spending is above that of Labour and in their 13 years they 'invested' heavily spaffing more money than the country could afford. We have high tax and high spend but it cant all be essential, we even have wartime levels of debt. The only way out of our debt is growth and yet calls are to spend more money and some justify it as more 'investment'.

[collinsl]

Re: Hmm

The absolute figures may be slightly higher, sure, however they should be much higher than they are to have kept up with a) inflation and 2) the increase in size of the population.

So in relative terms the cuts have been very deep, these have been partially hidden by economies and efficiencies found in new technology (when it's worked and hasn't been massively over budget and late) however we're now at the point where local councils are having to cut every service except that which is legally required, and some councils are stuck even there and are effectively bankrupt. Yes some of these are due to poor decisions and spending in the wrong places, but the vast majority are not, they are just down to the central government reducing funding per person served and not allowing council tax to go up enough to cope.

The police have suffered large scale cuts to their budgets per person served too, the whole thing from Boris Johnson's government about recruiting 20,000 extra officers in England and Wales only returns police officer numbers to just below those present in 2010, and doesn't account for most of the budget to pay those officers, which has had to come from police funds, meaning they've got rid of police staff, PCSOs, and have put warranted constables in to do their jobs instead, at higher pay rates than the police staff/PCSOs. Those then aren't officers on the streets where they can be effective at their jobs too.

Fire services have had to reduce the number of manned stations, relying more on volunteers and closing stations down completely, the NHS we're all aware of, services like Border Force are massively understaffed, the armed forces can't recruit enough people because their pay is too low, meaning their servicepeople get overworked and overstressed and leave, etc etc etc.

[codejunky]

Re: Hmm

@collinsl

"The absolute figures may be slightly higher, sure, however they should be much higher than they are to have kept up with a) inflation and 2) the increase in size of the population."

Not just absolute figures-

https://www.statista.com/statistics/298478/public-sector-expenditure-as-share-of-gdp-united-kingdom-uk/

"So in relative terms the cuts have been very deep"

How do we distinguish a cut from returning to normal? Remember the Labour gov massively 'invested' at unsustainable levels, selling gold and signing a lot of PFI contracts pre covid spending was finally coming down by about 2016/17 to labours blow out spending (before the global recession). That means the gov were reducing spending from the global crash highs, but still well above Blair/Browns splurging.

"some councils are stuck even there and are effectively bankrupt. Yes some of these are due to poor decisions and spending in the wrong places"

Things like councils investing in the markets and losing money during the great crash because they left money in Iceland banks (after being told to remove it). Or the vast 'investment' in 'renewables' which has harmed the economy all over. But this isnt just a councils issue they are all taking and spending too much. They are not only spending everything they take even during our high tax rates but also borrowing vastly more. The interest payments are at a 20 year high and growing-

https://www.bbc.co.uk/news/technology-67002195

While the rest of your comment concerns cuts in certain departments (I will assume you are right for now) the government has already spent the money. They are already spending too much and more. How is the NHS struggling after all that glorious unsustainable investment by Labour? Or the police?

"services like Border Force are massively understaffed"

I will single this out as it especially highlights the problem. The border force after massive investment and wondrous care was understaffed under Labour. At unsustainable levels of spaffing money they were understaffed. The legal knots to try and run a border brought us the stupid idea of high cost sending people to Rwanda and even that was tied up and messed with by activists before being dropped.

[Anonymous Coward]

Re: Hmm

It's so sad to see the enshittification of El Reg comment section by the same bot-like posting about the same tired old astro-turfed opinions on climate/economy/EU on EVERY SINGLE FUCKING ARTICLE, no matter the subject.

Back on topic.

How much money did the UK government "spaff" on the BBC's Computer Literacy Programme in the late 70s early 80s?

Plenty of people who comment on the pages of this august organ were products of this national investment.

Why is it when there is any small hint of dojng something similar, those who most likely previously profited from such endeavours always want to bar the next generation from having a similar leg up?

Anon. Because I've just had enough.

[Claverhouse]

The Road to Perdition

Barney Hussey-Yeo, founder and CEO of Cleo AI, took to X to criticize the move, warning that a reduction in investment and a raising of taxes risked pushing entrepreneurs to the US.

So the answer is to cut taxes and simultaneously increase investment from the government pot ?

Seems very Thatcherite.

The incoming administration

Not very British.