Chinese Company Develops 64-Core RISC-V CPU With American Company's Core Design as US Sanctions Loom

News
By
published

SophGo's SG2380 and SG2044 are based on SiFive's technology.

Data center
(Image credit: Shutterstock)

SophGo, a China-based chip developer, is working on several high performance processors featuring RISC-V-based IP that it licensed from SiFive, a major RISC-V IP designer from the U.S., reports HPCwire. This endeavor somewhat underscores the rising influence of RISC-V in the global chip industry and shows why some U.S. lawmakers are concerned about this open-source technology.

SophGo's first RISC-V-based project is the SG2380 processor, which includes 16 four-issue, out-of-order SiFive P670 cores, SiFive's X280 accelerator for AI/ML workloads, and Imagination Technologies' AXT-16-512 graphics processing unit. The CPU is mostly aimed at high-performance desktops, but it could also be used for edge servers that require 16 general-purpose cores as well as AI capabilities.

But the SG2380 is not the only SiFive-based product in SophGo's portfolio, as the company has already announced its SG2044 system-on-chip, slated for release in 2024. This 120W SoC will pack up to 64 high-performance RISC-V cores from SiFive, the final version of RISC-V vector extensions, PCIe 5.0, GbE, and LPDDR5x support. This processor will succeed little-known SGF2042, which supports 0.7 version of RISC-V vector extensions and is currently used mostly by researchers.

Both SophGo's SG2380 and SG2044 are to be produced on TSMC's 12nm-class process technologies. These products clearly exemplify the collaborative and innovative spirit of RISC-V, akin to the Linux operating system in its global contribution and development model. However, some U.S. lawmakers are consideration limiting RISC-V cooperation between American and Chinese companies. That sparked quite the controversy, with RISC-V International, headquartered in Switzerland, strongly opposing such governmental interference.

Despite the advancements and potential of RISC-V, its widespread adoption in the server and supercomputer sectors remains a distant goal. The current market is heavily dominated by x86 chips produced by Intel and AMD, with Arm also posing significant competition. Meanwhile, Arm and x86 technologies are controlled by companies based in the U.S. and U.K., which means that they are subject to export control regulations. As a result, Chinese developers are turning their attention to RISC-V, as they can either get open-source designs and expand on them, or just design high-performance technology from the ground up without any controls from the U.S. or U.K. governments.

Anton Shilov
Anton Shilov
Freelance News Writer

Anton Shilov is a Freelance News Writer at Tom’s Hardware US. Over the past couple of decades, he has covered everything from CPUs and GPUs to supercomputers and from modern process technologies and latest fab tools to high-tech industry trends.

See more CPUs News
14 Comments Comment from the forums
  • AndrewJacksonZA
    *dons tinfoil hat *
    So SiFive recently had layoffs, huh? Something-something roadmap?
    https://www.tomshardware.com/news/sifive-lays-off-hundreds-of-risc-v-developers
    Reply
  • TCA_ChinChin
    AndrewJacksonZA said:
    *dons tinfoil hat *
    So SiFive recently had layoffs, huh? Something-something roadmap?
    https://www.tomshardware.com/news/sifive-lays-off-hundreds-of-risc-v-developers
    I'm not too sure they are very related, although the possibility of sanctions would definitely hurt SiFive more. I think the factors that went into those layoffs happened earlier than when the US started considering risc-v sanctions.
    Reply
  • atomicWAR
    Open source is open source... governments need to get over it and stop trying to control it. Because the only other 'real' option is to outlaw or heavily regulate open source which would be a huge blow to the tech industry and slow down progress for all countries not just the 'targets' of such legal maneuvers. I am not saying sanctions don't have their place but open source needs to be off limits or it defeats the whole point of the open source community. But I suspect that weakening, dismantling or regulating the open source community may be the whole treacherous point, time will tell.
    Reply
  • Findecanor
    The article on HPCWire is misleading, and got misquoted.
    According to a SiFive employee on Reddit and Hacker News, the 64-core SG2044 will not have cores from SiFive but from the Alibaba-owned T-Head.

    The predecessor, SG2042 has 64 T-Head C920 cores. It is a variant of the T-Head C910.
    The C910 not only implements the open standard RISC-V, but the Chinese-developed core is open source. The C910 and C920 do however support only a draft version of the RISC-V Vector extension (needed for high-performance compute).
    The SG2044 is supposed to support Vector 1.0.

    SOPHGO has its roots in a spin-off of the Bitcoin operation Bitmain, so I think we should boycott them for that reason.
    Reply
  • bit_user
    This endeavor somewhat underscores the rising influence of RISC-V in the global chip industry and shows why some U.S. lawmakers are concerned about this open-source technology.
    @PaulAlcorn , would you please tell Anton it's an "open standard", not "open source"? If I had a nickle for every time I've seen this error, I could probably buy you a ticket to see the next Taylor and Travis show.
    ; )
    Reply
  • bit_user
    SophGo's first RISC-V-based project is the SG2380 processor, which includes 16 four-issue, out-of-order SiFive P670 cores, SiFive's X280 accelerator for AI/ML workloads, and Imagination Technologies' AXT-16-512 graphics processing unit.
    A few years ago, Imagination made a big announcement about throwing its hat into the RISC-V ring. If SophGo couldn't use SiFive cores, maybe Imagination would have something on par with the P670, before long. Or, maybe they'd license cores from Ventanna.
    https://www.tomshardware.com/pc-components/cpus/imagination-and-ventana-to-build-a-risc-v-cpu-gpu-platform
    The whole RISC-V cat is out of the bag. There are already non- US-based players in this market, and there'll only be more. Plenty of Chinese cores, too.
    Reply
  • bit_user
    AndrewJacksonZA said:
    *dons tinfoil hat *
    So SiFive recently had layoffs, huh? Something-something roadmap?
    https://www.tomshardware.com/news/sifive-lays-off-hundreds-of-risc-v-developers
    As I previously explained in the comments on that article, it doesn't necessarily mean SiFive is in trouble:
    https://forums.tomshardware.com/threads/sifive-lays-off-hundreds-of-risc-v-developers.3824334/post-23121908
    Reply
  • bit_user
    atomicWAR said:
    Open source is open source... governments need to get over it and stop trying to control it.
    Governments aren't trying to control it, as far as I'm aware. RISC-V isn't open source, either. It's an open standard, which is something different.

    What's interesting about RISC-V is that it should be far more resilient to government interference than ARM, not only because it's based in Switzerland, but also because it should be unencumbered by patents. Those patents are the legal basis by which the US has been controlling ARM and photolithography equipment.

    So, as the article points out, the US legislators are simply looking at restricting US companies from working with Chinese companies. They can't actually interfere with China's usage of RISC-V, itself.
    Reply
  • bit_user
    Findecanor said:
    The C910 not only implements the open standard RISC-V, but the Chinese-developed core is open source. The C910 and C920 do however support only a draft version of the RISC-V Vector extension (needed for high-performance compute).
    That's cool, but largely pointless. Other than potentially being able to implement one on a FPGA, being open source doesn't confer the same sorts of benefits as open source software.

    The problem is that building a commercially viable chip is such an expensive endeavor - and very little of that cost is addressed by the RTL design. Most of the NRE cost is incurred downstream of that. It also means when you get a chip from someone that allegedly contains an open source core, you can't verify it's exactly the same design as what you have the source code for. Therefore, it confers virtually none of the security benefits that people gain from open source software.

    FWIW, I don't see a repo for the C920 under T-head-Semi's github account. Also, looking at the activity graph for the C910, it appears they simply dumped it on github at the beginning of last year, rather than using github to actively develop it.
    Reply
  • atomicWAR
    bit_user said:
    Governments aren't trying to control it, as far as I'm aware. RISC-V isn't open source, either. It's an open standard, which is something different.

    What's interesting about RISC-V is that it should be far more resilient to government interference than ARM, not only because it's based in Switzerland, but also because it should be unencumbered by patents. Those patents are the legal basis by which the US has been controlling ARM and photolithography equipment.

    So, as the article points out, the US legislators are simply looking at restricting US companies from working with Chinese companies. They can't actually interfere with China's usage of RISC-V, itself.
    Regulation is a form of control So I don't get your bit of a rebutal over what is basically semantics. Of course US companies versions of risc v chips is a gray area.

    As for being open source vs open standard...I have seen both used A LOT interchangably for Risc V description in tech based reading/research. Yes open standard is more accurate imo but open source isn't entirely incorrect either as a reference point to risc v, especially for those techies where this area of expertise is not their specialilty (like me).

    I love your posts but every so often I don't quite see what your saying or conversely I don't agree. This is 'partly' one of those times. Though I suspect some it, knowing your detailed posts well, it's mostly about clarification...I hope lol as I'd hate to have a target on my back...mainly cause like me, you're tough to argue with as you'll source things into oblivion (again much like me). I do enjoy the back and forth though, always. This is how we learn, grow and become better more knowledgable techs/enthusiasts.
    Reply