3 ms·

RISC-V Foundation just moved[1] from the USA to Switzerland in late 2019 due to geopolitical concerns. Also, the European Union is heavily betting on the RISC-

• alexhutcheson 6y ago
  The foundation just coordinates the standards process and sets rules for using the trademark. It has a tiny budget and few staff. All substantive R&D is done by its members, which are companies of various size with headquarters in various countries: https://riscv.org/members/ https://riscv.org/members/ Changing where the foundation is incorporated doesn’t change where this R&D happens. Also Switzerland isn’t in the EU.
• GregarianChild 6y ago
  The RISC-V foundation moved to Switzerland primarily to help the big Chinese processor companies to work on RISC-V. Regarding the EU heavily betting on RISC-V, what gives you confidence that this will work? The history of EU funded projects to catch up with the US in AI, in cloud computing (like GAIA-X), in quantum computing, in the space race? HPC is not a mass-market, how should this niche be economically viable to amortise the huge cost of processor development? HPC in the US (and other countries) is often connected to military spending (although this is not always openly expressed). Processor development is extremely expensive, to the extent that Intel is now struggling to finance catching up with TSMC's fab process. Where is the EU's micro-architecture design capacity? ARM is about to be swallowed by Nvidia, and a lot of Arm's micro-architecture development is done in the US already. Now compare the micro-architecture design capacity with that in China (the world's biggest processor market). I really hope it will work out. I do too. Given 40 years of failure to catch up with the US, and now China, I am sceptical! I expect this to end up like previous EU attempts at catching up: the monies available will be much smaller than originally announced (by at least an order of magnitude), spent over much longer a time-frame, but more importantly, mostly flow to existing EU-based companies that are good ad lobbying, e.g. Siemens, and French conglomerates. Another prediction: China will soon dominate RISC-V, since they are now forced, by US sanctions, to move away from Arm. I cannot see big Chinese processor companies moving to local ISAs such as Loongson. Alibaba announced a RISC-V processor at ICSA this year [2]. The big US players (Intel, AMD, Nvidia) have much less of a need to move to RISC-V. [1] https://en.wikipedia.org/wiki/Loongson https://en.wikipedia.org/wiki/Loongson [2] https://conferences.computer.org/isca/pdfs/ISCA2020-4QlDegUf3fKiwUXfV0KdCm/466100a052/466100a052.pdf https://conferences.computer.org/isca/pdfs/ISCA2020-4QlDegUf...
  • nirv 6y ago
    > The RISC-V foundation moved to Switzerland primarily to help the big Chinese processor companies to work on RISC-V. Of course. It also removes unnecessary anxiety for the EU that the next ruling power in the US (or elsewhere) may eventually attempt to impose sanctions on the export of the IP. Switzerland, as part of Schengen and Union's single market, fits nearly perfectly into the narrative of European consolidation. > what gives you confidence that this will work? [...] Where is the EU's micro-architecture design capacity? I can't see the future, hence I have no confidence. But I express cautious optimism for the following reason: the issue of EU's dependence on some crucial technologies from the third parties is obvious, but there's no direct and quick solution to solve it. The EU can't just turn into venture capital Mecca overnight, the same holds for becoming a low tax heaven. However, the problems/challenges are there, and they must be addressed in the best way under the given circumstances. The EU has a well-established practice of funding R&D in large areas with non-specific goals or metrics, however this approach is not unique to the EU, it's crucial to the bloc to nourish the interbloc integration and establish logistics between the disparate parts of the market and institutions. It doesn't always lead to unexpected discoveries or breakthrough ventures, but in general it manages the integration task quite well. Which, in turn, contributes to the development of a rich ecosystem. As a tech entrepreneur in the EU, I'm not exactly happy with current affairs, but satisfied with the recognition of my industry's difficulties and proposed attempts to solve them. A failed venture with some new sprouts at worst case, a tangible breakthrough at best. So far, the EU holds leading position in the lithography systems (ASML). The EU is competitive in radio (Ericsson, Nokia) and embedded (NXP, Infineon, STM) semi manufacturing. There're many blank spaces on the European vertical technology stack, but it's far from a blank slate. This whole commitment is about EU pushing homegrown processor sovereignty across the entire stack. AFAIU, the project puts the focus on HPC as the only specific target platform that needs to be manufactured, but the developments around this project should lay the ecosystem for the boost of the industry as a whole. Companies such as Apple[1], Microsoft[2], Amazon[3], Google[3], and especially fresh start-ups like Nuvia[4] and Ampere[5] — are working on new ARM-based CPU designs, while SiFive[6] and some others on RISC-V. Also, I'd speculate that AMD and Intel will have to consider switching from the X86 in the foreseeable future. Now is the best possible time for venture into the semiconductor industry in the past few decades. I don't underestimate the complexity of the task. But perhaps a fraction of €145bn would be enough to plant a seed of a viable EU-based CPU semi enterprise; or is it definitely a doomed undertaking, eh? Maybe that French SiPearl[7] or some other company will take chances, we shall see. [1] https://en.wikipedia.org/wiki/Mac_transition_to_Apple_Silicon https://en.wikipedia.org/wiki/Mac_transition_to_Apple_Silico... [2] https://aws.amazon.com/ec2/graviton/ https://aws.amazon.com/ec2/graviton/ [3] https://arstechnica.com/gadgets/2020/04/google-wants-to-dump-qualcomm-launch-smartphone-soc-as-early-as-next-year/ https://arstechnica.com/gadgets/2020/04/google-wants-to-dump... [4] https://nuviainc.com/ https://nuviainc.com/ [5] https://www.anandtech.com/show/15575/amperes-altra-80-core-n1-soc-for-hyperscalers-against-rome-and-xeon https://www.anandtech.com/show/15575/amperes-altra-80-core-n... [6] https://www.reuters.com/article/us-sifive-tech-idUSKCN2571UL https://www.reuters.com/article/us-sifive-tech-idUSKCN2571UL [7] https://sipearl.com https://sipearl.com