Tuesday, February 18th 2020

UK Prepares $1.6 Billion for the Most Powerful Weather Forecasting Supercomputer

The UK government has set aside a budget of 1.2 billion GBP, which is roughly around 1.56 billion US Dollars. With this budget, the UK government plans to install the world's most powerful supercomputer used for weather forecasting in the year 2022. Previously, the UK government used three Cray XC40 supercomputers that are capable of achieving a maximum of 14 PetaFLOPs at its peak performance. The future system plans to take that number and make it look tiny. With plans to make it 20 times more powerful than the current machine, we can estimate that the future supercomputer will have above 200 PetaFLOPs of computing performance.

The supercomputer deployment will follow a series of cycles, where one is happening in 2022 and that supercomputer will be six times more powerful than the current solution. To get to that 20 times improvement, the supercomputer will get an upgrade over the next five years' time. While we do not know what will power the new machine, it will almost definitely be a CPU plus multi-GPU node configuration, as GPUs have gained a lot of traction in weather prediction models lately.
Add your own comment

43 Comments on UK Prepares $1.6 Billion for the Most Powerful Weather Forecasting Supercomputer

#1
john_
They are British, they know better, so I am expecting Xeons. /s
Posted on Reply
#3
remixedcat
Bet they still won't be able to be accurate enough to get current conditions correct let alone hours or days ahead lol
Posted on Reply
#4
phill
I really do hope it's an AMD built system.. That would be amazing :)
Posted on Reply
#5
notb
remixedcat
Bet they still won't be able to be accurate enough to get current conditions correct let alone hours or days ahead lol
What do you mean by "current conditions correct"? Current conditions are the input.
phill
I really do hope it's an AMD built system.. That would be amazing :)
Will it be less amazing with Intel inside?
john_
They are British, they know better, so I am expecting Xeons. /s
What's suddenly wrong with Xeons?
Posted on Reply
#6
XL-R8R
Seeing as this is England, it will either end costing £4.9 billion; or more likely, it'll never happen.




Check back in roughly 8-12 years when its complete/finally cancelled for the total cost.
Posted on Reply
#7
laszlo
just wonder what is the point of this forecasting as we have weather satellites; air mass movements can be predicted only for a short period after all...
Posted on Reply
#8
Imsochobo
notb
What do you mean by "current conditions correct"? Current conditions are the input.

Will it be less amazing with Intel inside?

What's suddenly wrong with Xeons?
What's wrong with xeons is that there is something A lot better for the task
nothing else.
Posted on Reply
#9
DeathtoGnomes
Soooo, Skynet wants to control the weather now?

notb
What's suddenly wrong with Xeons?
In a build like this, ya couple things come to mind... 1. Cost 2. Core Count per unit. 3. Cost....
Posted on Reply
#10
Basard
DeathtoGnomes
1. Cost 2. Core Count per unit. 3. Cost....
Taxpayer money... so definitely Xeons....
Posted on Reply
#11
Vya Domus
remixedcat
Bet they still won't be able to be accurate enough to get current conditions correct let alone hours or days ahead lol
It's just not true, weather forecasting is a ridiculously difficult thing to do and most predictions are quite accurate for such a complex phenomena. Given how variable and unpredictable the weather conditions are on small time intervals it's nothing short of amazing that they get it right the majority of time.
Posted on Reply
#12
Steevo
Look at how much we spent building it, programming it, feeding it data but it never matches reality, but it said we are all doomed so let's go with that.
Posted on Reply
#13
DeathtoGnomes
Basard
Taxpayer money... so definitely Xeons....
Budget build... so definitely EPYC....

10 Xeons = 15 EPYC and 100+ more cores = budget build. :p
Posted on Reply
#14
Tatty_One
Super Moder@tor
I think some of the reasoning behind this whether justified or not is a wish to get faster more detailed and localised predictions, for example, in the last 2 weeks or so we have had 2 huge storms coming in on the Jetstream, one a hurricane with months worth of rain in a day or two, huge parts of the country are currently flooded, most rivers throughout the country have burst their banks, many by up to 8ft of water, thousands homeless. Most of these places have flood defences, significant in some places but even they can't handle an extra 8 feet of water depth running in torrents, current forecasting says that areas of the country will get x amount of rain but we can't/don't break those areas down into small pieces and define within an area which towns/cities will get it the worst where some may not get it at all, not sure if this will work but it's in part the Government reacting to an outcry of home owners across the country asking what Government is going to do to make them safe as many of these homeowners cannot even get buildings or contents insurance because of where they live in "flood zones" and they cannot move because nobody wants to buy a house these days on a flood plain.
Posted on Reply
#15
Chrispy_
I'm expecting to see a lot of Nvidia silicon in the new Met Office Supercomputer. Tensor cores are absolutely ideal for particle simulation and fluid dynamics calculations.

I just wish Nvidia would stop dressing up Tensor cores as a gaming feature that requires developers to waste time on half-baked DLSS supersampling instead of adding content and/or polish to a new release.
Posted on Reply
#16
kurosagi01
Definitely going to cost more than 1.6 billion with the way how the UK approach things, also they can just continue reporting its going to rain everyday in the UK which is what we are known for having, don't need super computer for that.
Posted on Reply
#17
notb
laszlo
just wonder what is the point of this forecasting as we have weather satellites; air mass movements can be predicted only for a short period after all...
I can't believe you don't know why forecasting weather makes sense... Maybe that's not what you meant.

As for forecast quality:
Local forecasts (i.e. what you check to know if it's going to rain tomorrow) are very precise for 2-3 days ahead. This is not a computational limit - weather is just too random.
Large area forecasts (i.e. what will happen in your province/country as a whole - large pressure zones, hurricanes etc) can be modeled even few weeks ahead.
Imsochobo
What's wrong with xeons is that there is something A lot better for the task
nothing else.
There is nothing better. It's a cluster. You make it as big as you need.
What you actually mean is: cheaper. Based on EPYC it may be cheaper.

What you forget is that - even if EPYC's offer better value, they also offer less supply.
So not every supercomputer can be built with AMD interiors. This is one of the fundamental issues behind current Intel pricing (in all segments). 7nm is still relatively limited.
Posted on Reply
#18
phill
notb
Will it be less amazing with Intel inside?
What's suddenly wrong with Xeons?
It will have less of a core count and such, which I feel that if the work load is better suited to AMD, they'd be daft not to use them :) There's always options to try :)

kurosagi01
Definitely going to cost more than 1.6 billion with the way how the UK approach things, also they can just continue reporting its going to rain everyday in the UK which is what we are known for having, don't need super computer for that.
I wonder if that quote covers the 1Gb internet connection as well?? Or possibly 10Gb internet connection??...... :wtf::rolleyes::rolleyes:
Posted on Reply
#19
DeathtoGnomes
notb
What you forget is that - even if EPYC's offer better value, they also offer less supply.
So not every superComputer can be built with AMD interiors. This is one of the fundamental issues behind current Intel pricing (in all segments). 7nm is still relatively limited.
This assumes the pre-order depends on current inventory levels, which is does not, its in the pre-order stage still. Regardless of supply and chip costs, the EPYC is still the better value for such a build when it comes to the other factors that, i would think, matter more like power and core count. Things I hinted at already.
Posted on Reply
#20
remixedcat
notb
What do you mean by "current conditions correct"? Current conditions are the input.

Will it be less amazing with Intel inside?

What's suddenly wrong with Xeons?
All the computing power cant help if the weather algorithm is bad
Posted on Reply
#21
notb
phill
It will have less of a core count and such, which I feel that if the work load is better suited to AMD, they'd be daft not to use them :) There's always options to try :)
It's a cluster. You need 1000 cores, you buy 1000 cores.
AMD CPUs offer better value at the moment, i.e. you need less sockets to deliver the required cluster performance.

And of course there's a good chance this system can benefit from AVX-512. It really is a big selling point for Intel.
I wonder if that quote covers the 1Gb internet connection as well?? Or possibly 10Gb internet connection??...... :wtf::rolleyes::rolleyes:
First of all: speeds you've mentioned are for home use, not for datacenters...
Second: this supercomputer will likely be kept offline, i.e. no direct Internet access (it isn't needed).
Posted on Reply
#22
laszlo
notb
I can't believe you don't know why forecasting weather makes sense... Maybe that's not what you meant.

As for forecast quality:
Local forecasts (i.e. what you check to know if it's going to rain tomorrow) are very precise for 2-3 days ahead. This is not a computational limit - weather is just too random.
Large area forecasts (i.e. what will happen in your province/country as a whole - large pressure zones, hurricanes etc) can be modeled even few weeks ahead.

There is nothing better. It's a cluster. You make it as big as you need.
What you actually mean is: cheaper. Based on EPYC it may be cheaper.

What you forget is that - even if EPYC's offer better value, they also offer less supply.
So not every supercomputer can be built with AMD interiors. This is one of the fundamental issues behind current Intel pricing (in all segments). 7nm is still relatively limited.
i really can't imagine why this super computer is needed for forecasting; the satellites give us already the movement of air masses and the path also; sometimes is impossible to forecast in long term as a sudden change of air movement change all and no matter how powerful computer you use is not-predictable ; i'm sure the new one will forecast the same as the old one and both will be incorrect at long term forecast as you can't foresee how nature behave...
Posted on Reply
#23
xkm1948
For applications that can take full advantage of AVX512, Xeon remains a solid choice
Posted on Reply
#24
Chrispy_
phill
I wonder if that quote covers the 1Gb internet connection as well?? Or possibly 10Gb internet connection??...... :wtf::rolleyes::rolleyes:
Uh, the cost of internet that fast in the UK is negligible. 1Gb/s internet is typically £65/month for consumers, our business fibre (Daul 4Gb/s symmetric with failover to a different trunk) is £2700/quarter or £900/month. There is healthy fibre competition in Exeter where the Met office is based.

Assume that the power and cooling costs of a supercomputer cluster are 10X higher than the internet connectivity, and you'll be in the right ballpark at least.
Posted on Reply
#25
notb
remixedcat
All the computing power cant help if the weather algorithm is bad
Theoretically it's not that hard to write them, so - I'd imagine - there aren't many bad models in use by large (both national and private) entities.
It's PDEs after all. I mean: there are many lines of code, but the idea is quite simple.

Some fairly decent models / libraries are free and open-source. COAMPS is very popular:
https://www.nrlmry.navy.mil/coamps-web/web/home

Met Office uses a model called Unified Model. AFAIK it's not open-source, but it's fairly popular as well (I worked with it at university in Poland).
https://www.metoffice.gov.uk/research/climate/maps-and-data/uk-synoptic-and-climate-stations

As in many other problems in physics: it's actually the input data that makes this difficult - not the actual model complexity.
1) you're mostly using input from the surface, but you're actually modeling in 3D

2) the number of stations is limited. In case of UK and Met: just ~270 stations report in real time or hourly
https://www.metoffice.gov.uk/research/climate/maps-and-data/uk-synoptic-and-climate-stations

3) no matter how small your country is, you have to model the whole atmosphere anyway :)
This means you're forced to use 3rd party data for the rest of the planet. Of course national institutes cooperate (and some data is just free), but that means you have very little impact on data quality and data point density.

laszlo
i really can't imagine why this super computer is needed for forecasting; the satellites give us already the movement of air masses and the path also; sometimes is impossible to forecast in long term as a sudden change of air movement change all and no matter how powerful computer you use is not-predictable
I'm not sure where you're going with this. Numerical weather forecasting has been around for decades. It works.

We know equations that have proven to be very accurate. We come up with something better from time to time. It's not magic.

As for satellite imaging: you can't even learn the current conditions based on that, which implies that you can't make any kind of forecast.
Sure, they can be used an additional source, but there's very little data that can be extracted.
Posted on Reply
Add your own comment