Homebrew computer using RISC-V CPU?

Hello guys, I’m new and I’ve been keeping track of RISC-V since I heard of it last year or the year before. Now as far I as I can tell SiFive is the only company that has RISC-V hardware available for sell, and I’m wondering when the U54-MC CPUs will be available.

At this stage is it possible to build a Homebrew personal computer around RISC-V hardware. The free and open nature of RISC-V highly appeals to me as I’m a Linux user who no longer even uses Windows anymore.

Thanks for any answers .


SiFive have said 1st quarter 2018 for the U54-MC. With 1.5 GHz quad core that will be quite a usable PC. Hopefully it will be closer to an Odroid C2 than to a Raspberry Pi in performance.

For the moment if you want to run a real OS then you’re limited to a soft core in an FPGA and maybe 100 MHz. I’m old enough to remember when I’d have thought that was a pretty good PC, but it would be a bit painful running a web browser these days. It’s also going to cost a lot more than an ARM or x86 based machine.

The question of GPU also hasn’t been addressed. Those are fairly vital to the experienced performance of PCs these days.

Thanks, I was wondering about what GPUs were being looked at. Are there any free and open GPUs in development even on “paper”? If U54-MC boards are available with PCIe x16 slots I supposed we could use AMD Radon Rx 550 GPUs and higher since they have FOSS drivers available to examine.

BTW I’ve been building my own x86 computers since late 2000.

What do you define as “building”?

With all my x86 desktops since whenever the Athlon “Thunderbird” came out, which was … yeah … 2000, I’ve “build” them myself, defined as: pick a CPU, ATX motherboard with the same socket, RAM to fit the MB, disk with appropriate IDE/SATA/M.2 interface to suit the motherboard, case, power supply … and plugged it all together. A process taking an hour maybe.

In 1983 two friends and I built a 6809 computer by buying a 6809, some RAM chips, a ROM chip, some TTL logic chips, and we figured out which pins to connect to which pins and we wire-wrapped the whole thing. I was particularly proud of the clock circuit I designed which ran at 1 MHz while accessing the UART, at 2 MHz while accessing the RAM, and at 1,33 MHz (500 ns high, 250 ns low) when accessing the ROM as I noticed that although it was spec’d as 1 MHz it was only actually one half of the cycle that was critical.

We also wrote our own OS and compiler back end (from BCPL ocode).

There is quite a difference between the two versions of “building a computer”!


At the moment assembling out of already made parts, although I well aware such parts for RISC-V will be not available for awhile yet.

If you want to spend an interesting hour on Hackday it’s practically a sub-culture unto itself.

Thanks I’ll look into that.

Thanks again guys. I just got started with learning C yesterday using this book on Wikibooks.org

I figure that even if I don’t do any sort of programming, I can at least learn how to compile source code and help with bringing over more Linux distros for RISC-V such as Arch Linux and Manjaro Linux once RiSC-V hardware is available.

Worthy aspirations, good luck!

Thanks I found learning C and using GCC to compile the source files to be easier then I expected once I found a good tutorial herer:

The hardest part was simply getting started and sticking with it. On the plus side, I also getting better at using the CLI as well.