There is a project to have an EU designed processor for 2020. In the first phase it will be oriented to supercomputing.
After that they will move into other areas of computing.
Yes, future-accomplishments are always so amazing, aren't they? Especially since they aren't bound by actual metrics of success, real world challenges, compromises, faults, etc. Yawn. Let me know when it ships.
The project wasn't even considered necessary until relatively recently. It was born out of the need to achieve maximum protection for European users given that it was becoming impossible to offer real guarantees unless the entire thing was designed and built in Europe.
Of course it doesn't exist yet. That's what I said. I even said the initial version is not aiming to be a groundbreaker but simply a necessary first step. Since the original announcement, more countries have come onboard. Things will be coordinated from the Barcelona Super Computing Centre.
You can yawn as much as you want but the US stopped the Chinese from using Xeons in their supercomputers so they went ahead and designed their own processors (TaihuLight), jumping to first place in world rankings in the process. Could that have backfired harder for US business? Now the EU is going the same route but for different reasons.
Yawn away!
Yawn.
Nope, no backfire. The Sunway is essentially a shitload of SMIDs and narrowly focused for highly parallel tasks. It’s nothing like a Xeon and doesn’t translate into anything but...well making a shitload of SMIDs for some HPC domains. Impressive for what it is but we’re far better off with them on this rather than a Xeon+Phi platform that may benchmark in at less petaflops but can be applied to a wider range of problems.
The Chinese were already number one because they like to add a cabinet and claim the top spot. The prior Top was the tianhe2 which is a Xeon+Phi box.
The US is behind is due to budgeting so the Chinese and Swiss have the top 3 slots.
The ExaNoDe that you are crowing about is ARM v8 based. Which may be of interest to HPC folks if it works but pretty much a yawn for everyone else.
Its not going to challenge Intel, AMD or Apple.
And given the only advanced fab in the EU is owned by Intel those arm processors will be fabbed in Asia unless they are going with the 32nm process node (yawn) owned by STM.
A different view:
“It’s not based on an existing architecture. They built it themselves" ... "This is a system that has Chinese processors.” ... "This is the first time that the Chinese have more systems than the US, so that, I think, is a striking accomplishment,”.
Jack Dongarra, professor at the University of Tennessee (creator of the measurement method used by TOP500)
Being a shitload of [your technology of choice] is besides the point.
You missed the point again. The point is that they ever had to go to these lengths in the first place, not that they trounced the US in 2016 and 2017 (and look to continue their lead in 2018), which you try to justify because of a lack of funds - no one cares!
The EU will go the same route but basically for security reasons. Those reasons didn't exist (or didn't have as much importance) just ten years ago. Although the EU project will start out at the supercomputing level, it is planned to scale down to other applications. Some of the people involved speak of an 'Airbus' of supercomputing. Big yawn, right?. Who knows! I know the project will require billions in investment so it can only be pan European. No single member state has the finances to carry such a big project.
As for fabrication:
There is nothing to build yet!
do you really think no European company would be up to the task? And no, there would be no need to produce in Asia (in the near term).
Siemens, Infineon and Bull will probably be very interested in taking part.
Yes, I read the Dongarra paper. The difference is I understood it and you didn’t. They “trounced” the US in top supercomputer not because of any technological advantage but because the folks that buy supercomputers in the US haven’t done so since 2012.
You do realize that the US builds supercomputers not to be number one on a chart somewhere but to do stuff. Titan is 5 years old build using Opterons and K20s.
No, probably not.
The Chinese spent money and Oak Ridge and LLNL didn’t.
The bottom line is that Intel and IBM have a significant technical advantage over the Sunway. You can brute force your way forward and add enough cores to grab the top spot but it doesn’t mean anything in terms of applicable technology to mainstream computing.
You can hand wave the lack of EU fabs if you like but if you don’t build fabs overnight and it isn’t economically viable to do so just to make ARM chips for supercomputers.
And yes it’s a hugh Yawn. By the time the EU gets its shit together the Australians will have built a 10000 core Rasberry Pi cluster for LANL.
Dream on about billions of euros. The total budget allocated by the EU so far for exanode/mont blanc/whatever is 10.1 million euros.
Meanwhile Intel, Samsung, TSMC, etc spends $10-15 billion on fabs every major node and IBM invested $3B into quantum computing by itself.
Yeah, the EU is going to be a freaking supercomputing powerhouse because they have a massively underfunded R&D project.
You still don't get it.
There is little funding because the project is still being defined. The first real step was the declaration of March this year.
The EU plans (and ministers are serious when they say the final project could be in the same league as Airbus) draws from a number of other existing projects on HPC, funding of which has been approved in the hundreds of millions. Just read the link.
The exascale mission is clearly mentioned in the March declaration and when it comes to financing, the amounts set aside will be even higher than those already earmarked. There would be no way to reach the already aggressive timeframes laid out in the declaration without such funding.
Your 10.1 million euro claim doesn't mean anything as it is taken out of context.
I will go out on a limb and say if this project takes on an Airbus scale investment, you won't be yawning then.
As for fabs, do you even know what is available in Europe? I'd say you haven't got a clue.
There is a project to have an EU designed processor for 2020. In the first phase it will be oriented to supercomputing.
After that they will move into other areas of computing.
Yes, future-accomplishments are always so amazing, aren't they? Especially since they aren't bound by actual metrics of success, real world challenges, compromises, faults, etc. Yawn. Let me know when it ships.
The project wasn't even considered necessary until relatively recently. It was born out of the need to achieve maximum protection for European users given that it was becoming impossible to offer real guarantees unless the entire thing was designed and built in Europe.
Of course it doesn't exist yet. That's what I said. I even said the initial version is not aiming to be a groundbreaker but simply a necessary first step. Since the original announcement, more countries have come onboard. Things will be coordinated from the Barcelona Super Computing Centre.
You can yawn as much as you want but the US stopped the Chinese from using Xeons in their supercomputers so they went ahead and designed their own processors (TaihuLight), jumping to first place in world rankings in the process. Could that have backfired harder for US business? Now the EU is going the same route but for different reasons.
Yawn away!
Yawn.
Nope, no backfire. The Sunway is essentially a shitload of SMIDs and narrowly focused for highly parallel tasks. It’s nothing like a Xeon and doesn’t translate into anything but...well making a shitload of SMIDs for some HPC domains. Impressive for what it is but we’re far better off with them on this rather than a Xeon+Phi platform that may benchmark in at less petaflops but can be applied to a wider range of problems.
The Chinese were already number one because they like to add a cabinet and claim the top spot. The prior Top was the tianhe2 which is a Xeon+Phi box.
The US is behind is due to budgeting so the Chinese and Swiss have the top 3 slots.
The ExaNoDe that you are crowing about is ARM v8 based. Which may be of interest to HPC folks if it works but pretty much a yawn for everyone else.
Its not going to challenge Intel, AMD or Apple.
And given the only advanced fab in the EU is owned by Intel those arm processors will be fabbed in Asia unless they are going with the 32nm process node (yawn) owned by STM.
A different view:
“It’s not based on an existing architecture. They built it themselves" ... "This is a system that has Chinese processors.” ... "This is the first time that the Chinese have more systems than the US, so that, I think, is a striking accomplishment,”.
Jack Dongarra, professor at the University of Tennessee (creator of the measurement method used by TOP500)
Being a shitload of [your technology of choice] is besides the point.
You missed the point again. The point is that they ever had to go to these lengths in the first place, not that they trounced the US in 2016 and 2017 (and look to continue their lead in 2018), which you try to justify because of a lack of funds - no one cares!
The EU will go the same route but basically for security reasons. Those reasons didn't exist (or didn't have as much importance) just ten years ago. Although the EU project will start out at the supercomputing level, it is planned to scale down to other applications. Some of the people involved speak of an 'Airbus' of supercomputing. Big yawn, right?. Who knows! I know the project will require billions in investment so it can only be pan European. No single member state has the finances to carry such a big project.
As for fabrication:
There is nothing to build yet!
do you really think no European company would be up to the task? And no, there would be no need to produce in Asia (in the near term).
Siemens, Infineon and Bull will probably be very interested in taking part.
Yes, I read the Dongarra paper. The difference is I understood it and you didn’t. They “trounced” the US in top supercomputer not because of any technological advantage but because the folks that buy supercomputers in the US haven’t done so since 2012.
You do realize that the US builds supercomputers not to be number one on a chart somewhere but to do stuff. Titan is 5 years old build using Opterons and K20s.
No, probably not.
The Chinese spent money and Oak Ridge and LLNL didn’t.
The bottom line is that Intel and IBM have a significant technical advantage over the Sunway. You can brute force your way forward and add enough cores to grab the top spot but it doesn’t mean anything in terms of applicable technology to mainstream computing.
You can hand wave the lack of EU fabs if you like but if you don’t build fabs overnight and it isn’t economically viable to do so just to make ARM chips for supercomputers.
And yes it’s a hugh Yawn. By the time the EU gets its shit together the Australians will have built a 10000 core Rasberry Pi cluster for LANL.
Dream on about billions of euros. The total budget allocated by the EU so far for exanode/mont blanc/whatever is 10.1 million euros.
Meanwhile Intel, Samsung, TSMC, etc spends $10-15 billion on fabs every major node and IBM invested $3B into quantum computing by itself.
Yeah, the EU is going to be a freaking supercomputing powerhouse because they have a massively underfunded R&D project.
You still don't get it.
There is little funding because the project is still being defined. The first real step was the declaration of March this year.
The EU plans (and ministers are serious when they say the final project could be in the same league as Airbus) draws from a number of other existing projects on HPC, funding of which has been approved in the hundreds of millions. Just read the link.
The exascale mission is clearly mentioned in the March declaration and when it comes to financing, the amounts set aside will be even higher than those already earmarked. There would be no way to reach the already aggressive timeframes laid out in the declaration without such funding.
Your 10.1 million euro claim doesn't mean anything as it is taken out of context.
I will go out on a limb and say if this project takes on an Airbus scale investment, you won't be yawning then.
As for fabs, do you even know what is available in Europe? I'd say you haven't got a clue.
OMG. You really don’t know how this works.
The ministers are setting aside billions for the total initiative that covers everything from 5G technologies to climate change to space to computing. The money get piecemealed everywhere because everyone needs to feed at the trough.
When you look into the “read this before commenting” stuff, that you didn’t read, you will find that the allocation for H2020 FET Proactive - High Powered Computing is 4M euros for 2018 and 68M euros for 2019.
Bidding on stuff like this is what I used to do. Essentially it’s a huge jobs program for scientists and engineers in industry and academia scattered across member states (and for us congressional districts).
So let me clue you in.
When a politician says he’s getting you $200M for R&D what he means is he is getting you $150M for the research facility that will be built in his district, $25M for research grants for the local university and $25M for you...unless the building goes over budget. Which it will. After which you go back, hat in hand to try to get enough funding to finish whatever it was you wanted to do in the first place.
Expect to see some really nice HPC research centers appear because those are good for photo ops.
The U.S. could conceivably have claimed some of that money if Apple had decided to repatriate it, but the iPhone maker has so far avoided bringing cash back without a promise of lower taxes. Changes advocated by Republicans and the Trump administration could prompt Apple to take action.
...and IMO if they do the EU Commission won't care.
I think Vestager is honestly convinced she right in trying what she sees as leveling the playing field, attempting to rectify what she perceives as unfair advantages some big multinationals are claiming for themselves to the detriment of smaller companies and individual consumers. She's done so with Google, and she'll do so with Apple. Personally I believe she goes too far, seeing things that just don't exist on the level she thinks they do and applying corrective fixes with too heavy a hand, but I don't think it's just about the tax money from Apple which the EU doesn't get for themselves anyway. Any recovered taxes go to the the individual countries where the tax obligation originates.
The EU should care really. If Apple ends up losing in appeals, the U.S. could fire back at the EU by subjecting European companies to double taxation.
The US already does this... look up the 220% proposed tax hike against Bombardier's sales in the US.
The U.S. could conceivably have claimed some of that money if Apple had decided to repatriate it, but the iPhone maker has so far avoided bringing cash back without a promise of lower taxes. Changes advocated by Republicans and the Trump administration could prompt Apple to take action.
...and IMO if they do the EU Commission won't care.
I think Vestager is honestly convinced she right in trying what she sees as leveling the playing field, attempting to rectify what she perceives as unfair advantages some big multinationals are claiming for themselves to the detriment of smaller companies and individual consumers. She's done so with Google, and she'll do so with Apple. Personally I believe she goes too far, seeing things that just don't exist on the level she thinks they do and applying corrective fixes with too heavy a hand, but I don't think it's just about the tax money from Apple which the EU doesn't get for themselves anyway. Any recovered taxes go to the the individual countries where the tax obligation originates.
The EU should care really. If Apple ends up losing in appeals, the U.S. could fire back at the EU by subjecting European companies to double taxation.
The US already does this... look up the 220% proposed tax hike against Bombardier's sales in the US.
There is a project to have an EU designed processor for 2020. In the first phase it will be oriented to supercomputing.
After that they will move into other areas of computing.
Yes, future-accomplishments are always so amazing, aren't they? Especially since they aren't bound by actual metrics of success, real world challenges, compromises, faults, etc. Yawn. Let me know when it ships.
The project wasn't even considered necessary until relatively recently. It was born out of the need to achieve maximum protection for European users given that it was becoming impossible to offer real guarantees unless the entire thing was designed and built in Europe.
Of course it doesn't exist yet. That's what I said. I even said the initial version is not aiming to be a groundbreaker but simply a necessary first step. Since the original announcement, more countries have come onboard. Things will be coordinated from the Barcelona Super Computing Centre.
You can yawn as much as you want but the US stopped the Chinese from using Xeons in their supercomputers so they went ahead and designed their own processors (TaihuLight), jumping to first place in world rankings in the process. Could that have backfired harder for US business? Now the EU is going the same route but for different reasons.
Yawn away!
Yawn.
Nope, no backfire. The Sunway is essentially a shitload of SMIDs and narrowly focused for highly parallel tasks. It’s nothing like a Xeon and doesn’t translate into anything but...well making a shitload of SMIDs for some HPC domains. Impressive for what it is but we’re far better off with them on this rather than a Xeon+Phi platform that may benchmark in at less petaflops but can be applied to a wider range of problems.
The Chinese were already number one because they like to add a cabinet and claim the top spot. The prior Top was the tianhe2 which is a Xeon+Phi box.
The US is behind is due to budgeting so the Chinese and Swiss have the top 3 slots.
The ExaNoDe that you are crowing about is ARM v8 based. Which may be of interest to HPC folks if it works but pretty much a yawn for everyone else.
Its not going to challenge Intel, AMD or Apple.
And given the only advanced fab in the EU is owned by Intel those arm processors will be fabbed in Asia unless they are going with the 32nm process node (yawn) owned by STM.
A different view:
“It’s not based on an existing architecture. They built it themselves" ... "This is a system that has Chinese processors.” ... "This is the first time that the Chinese have more systems than the US, so that, I think, is a striking accomplishment,”.
Jack Dongarra, professor at the University of Tennessee (creator of the measurement method used by TOP500)
Being a shitload of [your technology of choice] is besides the point.
You missed the point again. The point is that they ever had to go to these lengths in the first place, not that they trounced the US in 2016 and 2017 (and look to continue their lead in 2018), which you try to justify because of a lack of funds - no one cares!
The EU will go the same route but basically for security reasons. Those reasons didn't exist (or didn't have as much importance) just ten years ago. Although the EU project will start out at the supercomputing level, it is planned to scale down to other applications. Some of the people involved speak of an 'Airbus' of supercomputing. Big yawn, right?. Who knows! I know the project will require billions in investment so it can only be pan European. No single member state has the finances to carry such a big project.
As for fabrication:
There is nothing to build yet!
do you really think no European company would be up to the task? And no, there would be no need to produce in Asia (in the near term).
Siemens, Infineon and Bull will probably be very interested in taking part.
Yes, I read the Dongarra paper. The difference is I understood it and you didn’t. They “trounced” the US in top supercomputer not because of any technological advantage but because the folks that buy supercomputers in the US haven’t done so since 2012.
You do realize that the US builds supercomputers not to be number one on a chart somewhere but to do stuff. Titan is 5 years old build using Opterons and K20s.
No, probably not.
The Chinese spent money and Oak Ridge and LLNL didn’t.
The bottom line is that Intel and IBM have a significant technical advantage over the Sunway. You can brute force your way forward and add enough cores to grab the top spot but it doesn’t mean anything in terms of applicable technology to mainstream computing.
You can hand wave the lack of EU fabs if you like but if you don’t build fabs overnight and it isn’t economically viable to do so just to make ARM chips for supercomputers.
And yes it’s a hugh Yawn. By the time the EU gets its shit together the Australians will have built a 10000 core Rasberry Pi cluster for LANL.
Dream on about billions of euros. The total budget allocated by the EU so far for exanode/mont blanc/whatever is 10.1 million euros.
Meanwhile Intel, Samsung, TSMC, etc spends $10-15 billion on fabs every major node and IBM invested $3B into quantum computing by itself.
Yeah, the EU is going to be a freaking supercomputing powerhouse because they have a massively underfunded R&D project.
You still don't get it.
There is little funding because the project is still being defined. The first real step was the declaration of March this year.
The EU plans (and ministers are serious when they say the final project could be in the same league as Airbus) draws from a number of other existing projects on HPC, funding of which has been approved in the hundreds of millions. Just read the link.
The exascale mission is clearly mentioned in the March declaration and when it comes to financing, the amounts set aside will be even higher than those already earmarked. There would be no way to reach the already aggressive timeframes laid out in the declaration without such funding.
Your 10.1 million euro claim doesn't mean anything as it is taken out of context.
I will go out on a limb and say if this project takes on an Airbus scale investment, you won't be yawning then.
As for fabs, do you even know what is available in Europe? I'd say you haven't got a clue.
OMG. You really don’t know how this works.
The ministers are setting aside billions for the total initiative that covers everything from 5G technologies to climate change to space to computing. The money get piecemealed everywhere because everyone needs to feed at the trough.
When you look into the “read this before commenting” stuff, that you didn’t read, you will find that the allocation for H2020 FET Proactive - High Powered Computing is 4M euros for 2018 and 68M euros for 2019.
Bidding on stuff like this is what I used to do. Essentially it’s a huge jobs program for scientists and engineers in industry and academia scattered across member states (and for us congressional districts).
So let me clue you in.
When a politician says he’s getting you $200M for R&D what he means is he is getting you $150M for the research facility that will be built in his district, $25M for research grants for the local university and $25M for you...unless the building goes over budget. Which it will. After which you go back, hat in hand to try to get enough funding to finish whatever it was you wanted to do in the first place.
Expect to see some really nice HPC research centers appear because those are good for photo ops.
No need to clue me in. LOL. I do contract work for employees at a Hungarian company whose sole business is obtaining EU grants for its customers in science and engineering.
Your idea of how things actually work bears little semblance to reality.
"...a European project of the size of Airbus in the 1990s and of Galileo in the 2000s..."
and
"This is a world class program that aims to increase EU computing capabilities by 100 times, the EuroEXA project is truly an exceptional collection of EU engineering excellence in this field,” said Peter Hopton, Founder of Iceotope and Dissemination Lead for EuroEXA. “We have all set our ultimate goal – to enable the power-efficient delivery of the world’s biggest supercomputer.”
As for you comments on financing, you are mistaken.
The hundreds of millions is on HPC. Within that we now have the Exascale project. I made that supremely clear.
Various exascale initiatives have come and gone over the years. Others are ongoing. The decision for some EU countries to band together on exascale computing was formalised in March 2017 with the formal declaration. Since then, many of the existing projects have been brought under the Exascale umbrella and initial funding has begun. More countries have come onboard.
Another aspect you seem not to grasp very well is that so far I have only mentioned the formal EU situation. On a national level, the different supercomputing centres also have their own national goals and business. There are nationally funded supercomputing initiatives that receive financing directly from national and regional governments. On top of course of the private funding and companies that also participate in these initiatives.
So, is the EU failing? No. Not bere or elsewhere. Latest polls seem to suggest that even the Brits regret Brexit!
China has moved (and is still moving) very fast and its roadmap is clear. The EU roadmap is very similar.
This article provides the necessary perspective and gives very clear pointers to where China wants to be in 5 years.
The EU plans (and ministers are serious when they say the final project could be in the same league as Airbus) draws from a number of other existing projects on HPC, funding of which has been approved in the hundreds of millions. Just read the link.
The exascale mission is clearly mentioned in the March declaration and when it comes to financing, the amounts set aside will be even higher than those already earmarked. There would be no way to reach the already aggressive timeframes laid out in the declaration without such funding.
Your 10.1 million euro claim doesn't mean anything as it is taken out of context.
I will go out on a limb and say if this project takes on an Airbus scale investment, you won't be yawning then.
As for fabs, do you even know what is available in Europe? I'd say you haven't got a clue.
OMG. You really don’t know how this works.
The ministers are setting aside billions for the total initiative that covers everything from 5G technologies to climate change to space to computing. The money get piecemealed everywhere because everyone needs to feed at the trough.
When you look into the “read this before commenting” stuff, that you didn’t read, you will find that the allocation for H2020 FET Proactive - High Powered Computing is 4M euros for 2018 and 68M euros for 2019.
Bidding on stuff like this is what I used to do. Essentially it’s a huge jobs program for scientists and engineers in industry and academia scattered across member states (and for us congressional districts).
So let me clue you in.
When a politician says he’s getting you $200M for R&D what he means is he is getting you $150M for the research facility that will be built in his district, $25M for research grants for the local university and $25M for you...unless the building goes over budget. Which it will. After which you go back, hat in hand to try to get enough funding to finish whatever it was you wanted to do in the first place.
Expect to see some really nice HPC research centers appear because those are good for photo ops.
No need to clue me in. LOL. I do contract work for employees at a Hungarian company whose sole business is obtaining EU grants for its customers in science and engineering.
Your idea of how things actually work bears little semblance to reality.
"...a European project of the size of Airbus in the 1990s and of Galileo in the 2000s..."
and
"This is a world class program that aims to increase EU computing capabilities by 100 times, the EuroEXA project is truly an exceptional collection of EU engineering excellence in this field,” said Peter Hopton, Founder of Iceotope and Dissemination Lead for EuroEXA. “We have all set our ultimate goal – to enable the power-efficient delivery of the world’s biggest supercomputer.”
As for you comments on financing, you are mistaken.
The hundreds of millions is on HPC. Within that we now have the Exascale project. I made that supremely clear.
Various exascale initiatives have come and gone over the years. Others are ongoing. The decision for some EU countries to band together on exascale computing was formalised in March 2017 with the formal declaration. Since then, many of the existing projects have been brought under the Exascale umbrella and initial funding has begun. More countries have come onboard.
Another aspect you seem not to grasp very well is that so far I have only mentioned the formal EU situation. On a national level, the different supercomputing centres also have their own national goals and business. There are nationally funded supercomputing initiatives that receive financing directly from national and regional governments. On top of course of the private funding and companies that also participate in these initiatives.
So, is the EU failing? No. Not bere or elsewhere. Latest polls seem to suggest that even the Brits regret Brexit!
China has moved (and is still moving) very fast and its roadmap is clear. The EU roadmap is very similar.
This article provides the necessary perspective and gives very clear pointers to where China wants to be in 5 years.
Do you doubt they will achieve their goals? And the EU?
I don’t doubt the Chinese but after watching Galileo flounder about for a decade I’m thinking eh.
Given the past history of Galileo, as in it was nearly dead a couple times, it’s all very well for EU ministers to pledge large future funding but this year and next isn’t billions.
And frankly, Galileo was more of a sop to the French than anything else. Ask Berry Smutny who got fired for speaking the truth and then outed by Wikileaks...
The difference between the E.U. and China in ability to execute is stark. The Brits may or may not regret Brexit but EU loses more than they do from the departure.
Especially considering ARM was UK until purchased by Japan...
Comments
There is little funding because the project is still being defined. The first real step was the declaration of March this year.
I suggest you read this before continuing:
https://ec.europa.eu/digital-single-market/en/high-performance-computing-contractual-public-private-partnership-hpc-cppp
The EU plans (and ministers are serious when they say the final project could be in the same league as Airbus) draws from a number of other existing projects on HPC, funding of which has been approved in the hundreds of millions. Just read the link.
The exascale mission is clearly mentioned in the March declaration and when it comes to financing, the amounts set aside will be even higher than those already earmarked. There would be no way to reach the already aggressive timeframes laid out in the declaration without such funding.
Your 10.1 million euro claim doesn't mean anything as it is taken out of context.
I will go out on a limb and say if this project takes on an Airbus scale investment, you won't be yawning then.
As for fabs, do you even know what is available in Europe? I'd say you haven't got a clue.
The ministers are setting aside billions for the total initiative that covers everything from 5G technologies to climate change to space to computing. The money get piecemealed everywhere because everyone needs to feed at the trough.
When you look into the “read this before commenting” stuff, that you didn’t read, you will find that the allocation for H2020 FET Proactive - High Powered Computing is 4M euros for 2018 and 68M euros for 2019.
http://ec.europa.eu/research/participants/data/ref/h2020/wp/2018-2020/main/h2020-wp1820-fet_en.pdf
Bidding on stuff like this is what I used to do. Essentially it’s a huge jobs program for scientists and engineers in industry and academia scattered across member states (and for us congressional districts).
So let me clue you in.
When a politician says he’s getting you $200M for R&D what he means is he is getting you $150M for the research facility that will be built in his district, $25M for research grants for the local university and $25M for you...unless the building goes over budget. Which it will. After which you go back, hat in hand to try to get enough funding to finish whatever it was you wanted to do in the first place.
Expect to see some really nice HPC research centers appear because those are good for photo ops.
http://money.cnn.com/2017/10/06/news/companies/boeing-bombardier-trade-ruling-tariff/index.html
Your idea of how things actually work bears little semblance to reality.
You really need to focus on this point:
https://insidehpc.com/2017/09/announcing-euroexa-project-exascale/
Two snippets from that piece:
"...a European project of the size of Airbus in the 1990s and of Galileo in the 2000s..."
and
"This is a world class program that aims to increase EU computing capabilities by 100 times, the EuroEXA project is truly an exceptional collection of EU engineering excellence in this field,” said Peter Hopton, Founder of Iceotope and Dissemination Lead for EuroEXA. “We have all set our ultimate goal – to enable the power-efficient delivery of the world’s biggest supercomputer.”
As for you comments on financing, you are mistaken.
The hundreds of millions is on HPC. Within that we now have the Exascale project. I made that supremely clear.
Various exascale initiatives have come and gone over the years. Others are ongoing. The decision for some EU countries to band together on exascale computing was formalised in March 2017 with the formal declaration. Since then, many of the existing projects have been brought under the Exascale umbrella and initial funding has begun. More countries have come onboard.
Another aspect you seem not to grasp very well is that so far I have only mentioned the formal EU situation. On a national level, the different supercomputing centres also have their own national goals and business. There are nationally funded supercomputing initiatives that receive financing directly from national and regional governments. On top of course of the private funding and companies that also participate in these initiatives.
So, is the EU failing? No. Not bere or elsewhere. Latest polls seem to suggest that even the Brits regret Brexit!
China has moved (and is still moving) very fast and its roadmap is clear. The EU roadmap is very similar.
This article provides the necessary perspective and gives very clear pointers to where China wants to be in 5 years.
https://www.top500.org/news/putting-the-rise-of-chinese-supercomputing-in-perspective/
Do you doubt they will achieve their goals? And the EU?
Given the past history of Galileo, as in it was nearly dead a couple times, it’s all very well for EU ministers to pledge large future funding but this year and next isn’t billions.
And frankly, Galileo was more of a sop to the French than anything else. Ask Berry Smutny who got fired for speaking the truth and then outed by Wikileaks...
The difference between the E.U. and China in ability to execute is stark. The Brits may or may not regret Brexit but EU loses more than they do from the departure.
Especially considering ARM was UK until purchased by Japan...