EU made CPU's? EU signs €145 billion declaration to develop next-gen processors and 2nm technology

[CarlBar]

1 minute ago, kuddlesworth9419 said:

The next 10-20 years are going to be really interesting in the CPU industry. 

 

 

May you live in interesting times is a curse for a reason.

[whm1974]

1 hour ago, CarlBar said:

 

May you live in interesting times is a curse for a reason.

More like a Mixed Blessing really.

[StDragon]

RISC-V is the future. It's the only clear path forward for any newbie into the semiconducting industry without paying a crapload of royalty fees (it is royalty-free). In the meantime, the world will get along just fine with a heterogeneous mixture of x86, ARM, and MIPS all intercommunicating at the software stack via APIs. But one by one, all will fallout of favor for RISC-V.

 

The only question I really have is the tipping point and curvature of adoption over time with regards to RISC-V. Its future is really a fete accompli.

[thechinchinsong]

1 hour ago, StDragon said:

RISC-V is the future. It's the only clear path forward for any newbie into the semiconducting industry without paying a crapload of royalty fees (it is royalty-free). In the meantime, the world will get along just fine with a heterogeneous mixture of x86, ARM, and MIPS all intercommunicating at the software stack via APIs. But one by one, all will fallout of favor for RISC-V.

 

The only question I really have is the tipping point and curvature of adoption over time with regards to RISC-V. Its future is really a fete accompli.

For what it's worth, I know some universities are switching or at least considering switching to a RISC-V curriculum for computer engineering and architecture away from things like ARM and MIPS. Probably won't matter until a decade or two down the line when many graduates have experience with RISC-V, but i guess we'll see how it pans out.

[Kroon]

12 hours ago, PeachGr said:

I think x86 licence can be bought at well

 

It will not be X86.  If you read the declaration it clear that the CPU will mainly be for embedded systems:

Quote

The signatory Member States agree to work together in order to bolster Europe’s electronics and embedded systems value chain. This will include a particular effort to reinforce the processor and semiconductor ecosystem and to expand industrial presence across the supply chain, in order to address key technological, security and societal challenges. We agree to consolidate and build on Europe’s position in areas of proven expertise, and aim to establish advanced European chip design capabilities and production facilities progressing towards leading-edge nodes for data processing and connectivity.

Source: https://ec.europa.eu/newsroom/dae/document.cfm?doc_id=71912

 

Exactly what architecture will be used are not decided yet, there 17 different wills in this mess and some of them want EU to develop a completely new architecture.  I cinserly hope they go for something like RISC-V, probably the only way since it suppose to be a free to use.

 

My guess is that this money will be gone before the research even starts.  To many countries want the money.  Companies like Nokia and Ericssson already approaching the governments.  This will all be a mess.

[CarlBar]

3 hours ago, StDragon said:

RISC-V is the future. It's the only clear path forward for any newbie into the semiconducting industry without paying a crapload of royalty fees (it is royalty-free). In the meantime, the world will get along just fine with a heterogeneous mixture of x86, ARM, and MIPS all intercommunicating at the software stack via APIs. But one by one, all will fallout of favor for RISC-V.

 

The only question I really have is the tipping point and curvature of adoption over time with regards to RISC-V. Its future is really a fete accompli.

 

Couple of threads going on so not sure which thread i covered this in previously, but as i pointed out in that post, corporate will not move from x86 unless given absolutely no other choice, and what they use dictates what OS users are most familiar with, (and need to be taught in school), what most software developers develop for, and thus whats most popular at home.

 

RISC isn;t going to stop being niche ever unless Microsoft goes under or decides to alienate all it's biggest customers by ceasing x86 support. Even if ti wanted to, in the current world witout intel jumping on boards it's DoA. there isn't the enough high performance CPU fabbing capability without intel right now. And even if intel jumps on board, if they band together AMD and NVIDIA could really make it hard to move if they don't support it GPU side.

 

In the real world any move away from x86 requires the joint efforts of multiple hardware and software vendors all actively pushing a specific change along in unison. And none of them have any vested interest in doing so. ARM got a foot in the door because smart phones by their nature where and had to be a complete break from desktop, they weren't targeted at most of the use cases corporate cares about, they wouldn't work with a desktop like interface, and there was no pre-existing standard to get in the way. It let thenm walk around the backwards compatibility and familiarity barriers.

[StDragon]

11 minutes ago, CarlBar said:

 

Couple of threads going on so not sure which thread i covered this in previously, but as i pointed out in that post, corporate will not move from x86 unless given absolutely no other choice, and what they use dictates what OS users are most familiar with, (and need to be taught in school), what most software developers develop for, and thus whats most popular at home.

 

RISC isn;t going to stop being niche ever unless Microsoft goes under or decides to alienate all it's biggest customers by ceasing x86 support. Even if ti wanted to, in the current world witout intel jumping on boards it's DoA. there isn't the enough high performance CPU fabbing capability without intel right now. And even if intel jumps on board, if they band together AMD and NVIDIA could really make it hard to move if they don't support it GPU side.

 

In the real world any move away from x86 requires the joint efforts of multiple hardware and software vendors all actively pushing a specific change along in unison. And none of them have any vested interest in doing so. ARM got a foot in the door because smart phones by their nature where and had to be a complete break from desktop, they weren't targeted at most of the use cases corporate cares about, they wouldn't work with a desktop like interface, and there was no pre-existing standard to get in the way. It let thenm walk around the backwards compatibility and familiarity barriers.

The end-game is "ambient computing", where your online profile follows your physical and virtual presence. Most of your interaction via augmentation will occur through hardware you won't own; hence ambient.

 

Right now there's a big push to cloud everything and stitch it all together with API. It might take another 15 to 20 years to phase out x86 to irrelevancy, but it will happen nonetheless.

[CarlBar]

3 minutes ago, StDragon said:

The end-game is "ambient computing", where your online profile follows your physical and virtual presence. Most of your interaction via augmentation will occur through hardware you won't own; hence ambient.

 

Right now there's a big push to cloud everything and stitch it all together with API. It might take another 15 to 20 years to phase out x86 to irrelevancy, but it will happen nonetheless.

 

And that affects what i said how. You can move everything into the cloud, (i guarantee most major businesses won't have for security alone), thats not going to stop businesses demanding extremely heavy control over whats going on with their stuff there and moving their existing software there, all of which will be x86 so the cloud will be x86.

 

There's a reason Apple has next to no presence in corporate space, (relatively speaking). It;s because apples super locked down, everything obsolete eventually, you have no control style, is something corporate space absolutely will not accept. So they told apple where to shove it.

 

Thats the issue with killing x86. Corporations will not let it die, it's too problematic for them to let it. They also want reliability and control. Any implementation of x86 that doesn't give them that won't get anywhere.And whilever the major corporations are using it it will persist in the home space because it's what people are familiar with and will be taught in school.

[kaixax555]

Would be interesting, and I would suspect a lot of innovation would be driven around Dutch companies.

A quick breakdown on some of the Dutch semiconductor/semiconductor equipment manufacturers.
ASML - Probably one of the most advanced companies out there right now, originally a joint venture between ASM International (where it got its name) and Phillips. They are currently at the forefront of EUV lithography, and would likely dictate the feature size of future chips.

ASM International - As above, originally one half of the joint venture ASML before it was an independent company. Produces equipment for front end of chip manufacturing (mainly material deposition). Also owns 25% of Hong Kong/Singapore based ASM Pacific Technology, which produces equipment for back end of chip manufacturing (wire/die bonding, SMT, packaging, AOI)

NXP - Spun off from Phillips, fifth largest non-memory semiconductor manufacturer, mainly dealing with automotive, secure identification and digital networking. Also inventor of Near Field Communication (NFC) technology. Has 7 wafer fabs and 7 test and assembly sites around the world.

 

Besi - Fast rising semiconductor equipment manufacturer, mainly produces equipment for backend (die attach, leadframe, packaging).

There are others in Europe that I have not listed, but I think it will surprise a lot of people that Netherlands has a surprisingly large presence in the semiconductor space, especially in the area of semiconductor equipment. American companies still dominate most of the space (Applied Materials, KLA, Lam Research, Teradyne, etc.) but I wouldn't be surprised if the Europeans can match up.

[McBUrn]

That's an interesting news and with the power/money/human resources in Europe this could very well succeed like it has in other industries in the past.

However as always in the EU, everybody wants the money and the jobs on their own soil. That's not a federal state so those discussions tend to overcome the ones that actually matter for the said project and end up killing it. Guess there is already a massive lobbying from various companies to get the money and countries to get the fabs ...

[Mihle]

RISC V have a chance to compete with ARM, but it doesn't with x86 unless it ends up being multiple times the performance, and even then it would take many many years after that, and software companies would be the ones deciding it.