A major Russian company is testdriving Arm-based PCs.
Russian Nuclear Company Tests ‘Beaver’ PCs With Homegrown Baikal CPUs : Read more
Russian Nuclear Company Tests ‘Beaver’ PCs With Homegrown Baikal CPUs : Read more
You looked at the specs, right? My phone from 2015 had ARM A57 cores, and I think they ran at a higher frequency than this SoC. They would almost be better off just buying some 8 GB Raspberry Pi 4's from the grey market ...except for the lack of ECC memory and RAS features that you'd want, if you're going to use them in a nuclear plant.These are perfect all in one systems for light weight use.
You looked at the specs, right? My phone from 2015 had ARM A57 cores, and I think they ran at a higher frequency than this SoC. They would almost be better off just buying some 8 GB Raspberry Pi 4's from the grey market ...except for the lack of ECC memory and RAS features that you'd want, if you're going to use them in a nuclear plant.
Okay, but you said:It's not about speed or specs;
Baikal-M1 was theoretically interesting as a Raspberry Pi 4 desktop alternative (slower single, faster multi, much faster Mali GPU*) in a universe where they weren't ultra expensive and annihilated by sanctions. And a higher cost could have been justified by the feature set (RAM support, I/O) and Russian security concerns.You looked at the specs, right? My phone from 2015 had ARM A57 cores, and I think they ran at a higher frequency than this SoC. They would almost be better off just buying some 8 GB Raspberry Pi 4's from the grey market ...except for the lack of ECC memory and RAS features that you'd want, if you're going to use them in a nuclear plant.
Hmmm... from what I've read, the open source driver for Mali GPUs is in quite a dismal state. And if you have to use the proprietary driver, well... let's say that's not great for security.*Just based on GFlops reported for VideoCore VI and this MP6 version of the Mali-T628, it could be over 4x faster. Mali is also better supported than VideoCore. And of course, there's the PCIe 3.0 x8 capability for a discrete GPU.