Not sure how they compare exactly. And of course it never hurts to have two players in the game.
I just meant I've heard a lot more about stuff going on in the RISC-V field in China. A lot of it is focused on the embedded stuff yes but not only there.
Yup. Their best builders run JH7110 (based on the in-order SiFive U74 cpu, RVA20 + partial B) and these are slow, somewhere between rpi3 and rpi4.
This will improve in 2026, with the first chips integrating RVA23 microarchitectures, such as the Tenstorrent Atlantis SoC and development board, with Ascalon, announced for 2026Q2.
Are there RISC-V chips with comparable performance for desktops and laptops on the market?
I thought so, yes: https://www.theregister.com/2025/03/05/china_alibaba_risc_v_...
Not sure how they compare exactly. And of course it never hurts to have two players in the game.
I just meant I've heard a lot more about stuff going on in the RISC-V field in China. A lot of it is focused on the embedded stuff yes but not only there.
Are C930 CPUs actually available?
For the Loongson 3A6000, there is an independent performance review: https://chipsandcheese.com/p/loongson-3a6000-a-star-among-ch...
I couldn't find something comparable for C930 CPUs.
Debian's RISC-V machines are rather slow.
It takes several days to build the gcc-15 package in riscv64 but just a few hours on loong64.
Yup. Their best builders run JH7110 (based on the in-order SiFive U74 cpu, RVA20 + partial B) and these are slow, somewhere between rpi3 and rpi4.
This will improve in 2026, with the first chips integrating RVA23 microarchitectures, such as the Tenstorrent Atlantis SoC and development board, with Ascalon, announced for 2026Q2.