China to publish policy to boost RISC-V chip use nationwide, sources say https://www.reuters.com/technology/china-publish-policy-boos...
I don’t see anything but regret for Europe several decades from now if they decide to start providing China with the technical expertise they’re currently lacking in this space.
This is all about China trying to find a way to escape the pressure of sanctions from Europe and the US.
Not as friends or allies, but there aren't a lot of those left anyway. It's only rational in this multi polar world to have some level of engagement with all parties.
Most of the sanctions Europe have on China were just to please the US anyway.
We are reliant on the US as only 2 companies can make the x86/64 chips. I don't think Europe would be completely against working with a US or Chinese company like Hi Five/Star Five, as long as we weren't dependent on them, and could pull ties if they abused their position of control.
might as well collaborate with them on this as well.
I know the US get to have elections but its always between family dynasties, billionaires or corporate stooges. The choice is an illusion.
Lets cut out the middleman and know we're working with a different system for societal structure rather than one that pretends otherwise.
We need these cheap-ish computers in the hands of people who will port software to the platform. Without a good selection of ready to go software, the hardware is pretty irrelevant.
https://sifive.com/boards/hifive-premier-p550
https://sifive.com/boards/hifive1-rev-b
https://www.beagleboard.org/boards/beaglev-fire
https://starfivetech.com/en/site/boards
https://milkv.io/docs/duo/getting-started/duo256m
https://www.raspberrypi.com/products/rp2350/The biggest problem with them is software. Many boards only have buildroot SDKs or niche and outdated. Fedora related images.
Though if you're experienced you can port your own linux distributions to these boards.
I would currently recommend the BananaPI BPI-F3 or the OrangePI RV2 for that purpose, as they both have the same SpacemiT X60 cores, which support the vector extension.
Sadly there are currently only in-order cores with RVV support available. Getting a cheap out-of-order implementation is the next most important thing for improving software support.
Tenstorrent has announced they will release a 8x Ascalon devboard and laptop next year: https://youtu.be/ttQtC1dQqwo?t=1035
Discussion here: https://news.ycombinator.com/item?id=43309376
Gives you something to play around with, very inexpensively.
But really, virtual machines may be preferable; at least to get started.
And what shall one emulate in a VM ? A nonexisting physical processor ? /s
I know there is Orange Pi Riscv but maybe there are other cheap hardwares.
Yeah they'll be slow but nothing can be slower than an x86 loaded with a Windows 11 or something on it.
I nearly spit out my tea laughing.
That's going to be hard to beat, yes..
So youd have to convince a lot of people to learn other tools
> The first phase of this six-year endeavor is backed by €240 million (£200 million, $260 million) in funding.
For this to be a serious effort it would take another two zeros at the end of that number. This is 100x too small.
In 6 years, we'll have spent a pittance, to realize that we got basically nothing for it, and we're even further behind the US whose companies are spending tens of billions to develop new accelerators.
Let's take one US company at random, Groq, they've raised 10x this amount of money. That's one startup. Never mind Cerebras, SambaNova, Tenstorrent, etc. How is this effort going to compete? And they're giving the money to "38 leading partners" instead of one focused entity. It won't compete. It's just a waste.
The EU is still thinking too small. In an era where the US is no longer a reliable partner (maybe even a rival), and where Taiwan could disappear overnight, this is extremely stupid and dangerous.
I don't understand why the EU can't get serious about tech. Why does every investment need to be peanuts? Why can't we pay people well so they don't all leave to the US/Canada? Why can't we seriously invest in startups?
Private investors in Europe don't have the very deep pockets of US tech investors and there is much less of a culture of risk taking in investing in Europe on top of that
Edit: to be clear, I agree with your general point.
Maybe someone from Europe could weigh in? I'm probably wrong, if the funding is transparent it should be easy to confirm or deny.
1. It is a single entity but composed of teams from 38 different partners. They have a "consortium". It has its disadvantages but it is not a completely independent funding.
2. The "consortium" may have asked for much more (between 5x to 20x more) and was probably denied.
As to your question of why we cannot seriously invest in startups?
- Because we do not have a single funding agency across whole of Europe. Each country has its own funding problems and Germany has its debt brake. So, the funding is not unified. Each country wants to fund 10 of its own startups instead of Europe by itself funding 10. This means 10x less money. EU Horizon projects didn't focus on industry at all. EuroHPC is a very new, 4 years into its first projects.
- There's no funding because we do not have customers! None of European tech companies will benefit from chips enough to invest in new tech. All of them are running old technologies. Car companies are to blame here because they are the biggest customers in Germany and they think of themselves as only car companies. No one in Germany is doing AI for cars for FSD for example. In general, European consumption is very backward and low-tech.
- Europe is finding it very hard to raise capital from outside Europe due to various reasons. Like Groq raised 650M from Saudi Arabia. In Europe, that is politically impossible.
The US has been funded by an insane level of debt for the last 60+ years. Debt that might come calling quite soon, according to Donald Trump's own treasure secretary, iirc, and might even be the reason for all the current apparent international Trump-craziness (well, Trump being a narcissist certainly doesn't help).
While the EU has serious debt, if I understand correctly, that's several orders of magnitude smaller when compared to GNP, which limits the ability of the EU to invest.
I also found this report on their FPGA Emulation Platform: https://www.riser-project.eu/wp-content/uploads/2024/11/RISE...
So from these resources it seems like they develop a vector processor with Semidynamics out-of-order Atrevido core as a scalar core and their Vitruvius VPU.
There is a paper about a previous iteration of the VPU: https://dl.acm.org/doi/10.1145/3575861
In the more recent report they have a vector length of 16,384 bits, with 16 lanes (8 in FPGA, 16 in the diagram, final version could be more), so total of 16*64=1024 bits of ALUs.
Slide 15 seems to indicate that they want to create a chip with 32 of those cores, a shared L3 cache, and access to HBM.
- Who Owns it (Japan) - Where is it headquartered (Cambridge) - Where is most of the IP produced (Cambridge mostly, but the remainder is in the US)
So if we care about being fast, surely the most expedient way, complete with guaranteed success, is to simply buy out softbank and then bring any IP development that's been offshored to the US back to Europe?
It is pretty hard to call Arm, or any decently sized modern company, 100% anything.
Besides they are no longer 100% as you mention.
> Europe bets on supercomputing sovereignty
I'm laughing and dying inside. Europe has forfeited all possibilities on creating their own chips, partially because of production regime, partially because we were never good in this subject. At the moment the war is already lost (yes, I consider any negotiation for resources a war, whether in law creation or in movement of forces). Therefore, we're condemned to rely on China's supplies, and chips supplied by China will have this
https://www.techspot.com/news/107073-researchers-uncover-hid...
which defies idea of sovereignty at all.
the war against whom?
And chips are everywhere. Hence, the EU is dependent on China. Q.E.D.
Or maybe I am mistaken. Please tell me what you think regarding the above.
This seems very focused on current architecture, which could be replaced with something more novel without the fundamental limits of matrix multiplication.
China is pushing RISC V aggressively, and might be a lot more likely to succeed in making competitively powerful cores than €240M pounds spent in Europe, where money won’t go nearly as far.
I imagine one of the biggest constraints on success here is just expertise. If Apple’s hardware team, or Qualcomm’s Oryon team were tasked with making a high performance RISC V CPU, I’m sure they could crank out something incredible pretty quick. But I have a feeling practical expertise on this sort of cutting edge hardware design is a rare thing. Frankly no idea how this human capital compares between Europe and China, but I’ll be excited to see progress and genuine competition on open architectures like this
The _current_ big EU supercomputer initiative does use ARM designs (https://www.theregister.com/2024/05/14/sipearl_rhea1_specs/ ), but you wouldn’t necessarily want to be totally dependent on them if you can help it.
ARM just announced they are manufacturing their own chips for the first time further threatening their customers (despite testifying the exact opposite in court a couple months ago).
Since SoftBank took over, their company has shifted and proved that when a standard is controlled by one company, there will eventually be issues.
Switching to RISC means those issues won’t ever happen again.
RISC-V being based out of Switzerland, the ISA being under a permissive Creative Commons license, and most software tools being FOSS is definitely why it's being adopted here. It's completely isolated from all geopolitics.
The US (certainly the current US) can do that, but (https://en.wikipedia.org/wiki/Arm_Holdings):
“Arm Holdings plc (formerly an acronym for Advanced RISC Machines and originally Acorn RISC Machine) is a Japanese-owned British semiconductor and software design company based in Cambridge, England”
The world is abandoning rent seekers.
ARM will get you power faster.
But in the long run, RISC-V will be the most powerful.
Even more, some considerations could have more weight then architecture for particular case.
Examples are good compiler/libs/frameworks, some specific software, good support, experience on similar contracts, big number of professionals with military clearance.
That's why some long time IBM won most govt contracts on supercomputers.
But once IBM decided, govt is not interest enough client and after that moment, most contracts won by Intel.
Examples was video chips, io chips, MMU, numerical coprocessor.
At that time (8086) Intel produced even RAM and ROM chips, so govt could buy all from one contractor, and this is also good in some cases.