Developerbox (Possible High-End RISC machine?)

Hi I know it has been mentioned briefly on these forums before but does anyone know if RISC OS could be easily ported to the Developerbox?

See link here: DeveloperBox

Could this be one of the most powerful RISC OS machines potentially as a nice desktop PC bundle?

The “Hardware Documentation” page contains virtually no documentation so we’re not off to a good start…

If you want to port it yourself, welcome!

But as Chris says , it is impossible without real hw docs!

I believe this is one of those “large number of low power/performance cores” CPUs. Without fully multi-threaded OS and apps (and sometimes even with!), I suspect current systems would actually be faster. In some case, I suspect even Pi3 would be faster.

That being said, the feature set of the board is nice, it is just a shame the CPU would probably be a bit insipid for the cost, and the development/porting work would be substantial.

Being blunt, I’m not sure what this offers a RISC OS user that one of our Titans (similar form factor etc) does not.

AArch64…

AArch64 might not a problem. The Socionext SC2A11 is based on a Cortex A53 which, if I remember correctly, still has a proper AArch32 mode.

AArch64 might not a problem. The Socionext SC2A11 is based on a Cortex A53 which, if I remember correctly, still has a proper AArch32 mode.

Chris is right. The whole thing is academic if there’s no real hardware documentation.
So, potentially a lot of research work to try and get a decent set of documents and end up with a spec that’s not as useful as existing units.

Unless someone wants the experience of porting to a new hardware set you have to say why not buy a Wandboard and a case with PSU and use the iMX6 beta to experiment with?
If you want a finished product then R-Comp have a range for you to choose from.

As a porting candidate or as a developer’s box SolidRun’s HoneyComb ARM workstation might be worth a look, especially in terms of cost and performance – #HoneyComb_LX2K . ;-) And the pre-order price seems attractive – $550 …

As a porting candidate or as a developer’s box SolidRun’s HoneyComb ARM workstation might be worth a look, especially in terms of cost and performance – #HoneyComb_LX2K . ;-) And the pre-order price seems attractive – $550 …

Most of it cannot easily be used by RISC OS, e.g. the vast amount of RAM and the 16 cores, which just means that 15 of them are currently unused…and you’d need a graphics card with drivers over PCIe.

Once there are USB3 drivers, a RPi 4 gives you nearly the max RISC OS will currently be able to support (minus S-ATA). For $55 (for the 4 GiB variant). Which seems to be a lot more attractive. In the meantime, the Titanium should suffice for S-ATA fans :-)

Most of it cannot easily be used by RISC OS, e.g. the vast amount of RAM and the 16 cores, which just means that 15 of them are currently unused…and you’d need a graphics card with drivers over PCIe.

Sorry to be a wet blanket on this sort of idea but I’ve either been looking in the wrong place or there isn’t any meaningful documentation.
If there was any documentation beyond a long specification sheet it might be possible to decide whether it was even worth thinking about.
It would appear there is not. Case closed.

To modify my earlier statement:
If you want a product that definitely runs RISC OS then R-Comp and CJE have a range for you to choose from, or for the DIY version buy a Wandboard or Titanium and use the downloads from the ROOL site.

BTW for the price quoted (ignoring shipping costs etc.) I believe you could buy a lifetime support package from R-Comp and a couple of Wandboard-Quad units where RO is known to work already.
That DIY option works for people anywhere on the globe AFAIK.

Most of it cannot easily be used by RISC OS, e.g. the vast amount of RAM and the 16 cores, which just means that 15 of them are currently unused…and you’d need a graphics card with drivers over PCIe.

Sorry to be a wet blanket on this sort of idea

Correct me if I’m wrong, but I think Steffen was also firmly in the wet blanket category.

You say it’s a bad choice because there’s no documentation. He says it’s a bad choice because for the foreseeable future we’ll only be able to use 1% of its potential (and potentially lack of documentation for graphics cards).