21 comments

[ 3.7 ms ] story [ 62.8 ms ] thread
It would be cool to see how well you could make a Godson-3A run JS, some of their instructions designed to help virtualization could help (I guess).

A bit random that this showed up today, was just thinking about MIPS. Partially SGI nostalgia, and partially about the new Loongson/Godson-3B chips and their fancy vector units.

Why?
Because MIPS needs a good, high-performance JavaScript engine?
Because there is a very significant amount of devices running MIPS processors, including a bit of upcoming Android stuff,.
The Chinese are aggressively building new MIPS designs (called Loongsoon/Godson) and building them into hardware which is totally open (see RMS's Lemote Yeelong). Having a fast JS engine for MIPS could help these systems become more viable for those who aren't staunch Openness gurus like RMS.

There's also a large community of users of old SGI machines, and a few of them have a good amount of free time to pursue MIPS projects like this.

Are schematics available? (A quick search didn't turn any up for me.) I don't think I'd call a product totally open without them.
Schematics for the chips are not available, but you can boot them without any non-free software or firmware. Which is why RMS prefers the platform coupled with gNewSense.
The parent poster meant that they can run only free software, including BIOS and firmware.

Stallman:

"Because copying hardware is so hard, the question of whether we're allowed to do it is not vitally important. I see no social imperative for free hardware designs like the imperative for free software. Freedom to copy software is an important right because it is easy now--any computer user can do it. Freedom to copy hardware is not as important, because copying hardware is hard to do. Present-day chip and board fabrication technology resembles the printing press. Copying hardware is as difficult as copying books was in the age of the printing press, or more so. So the ethical issue of copying hardware is more like the ethical issue of copying books 50 years ago, than like the issue of copying software today."

http://www.linuxtoday.com/news_story.php3?ltsn=1999-06-22-00...

If totally open only means the software, aren't some x86 boxes there already, assuming you can replace the BIOS with coreboot?

My point on schematics is this: I don't want schematics for consumer products so that I can copy the hardware; I want schematics so I can troubleshoot and learn. Maybe that's just me (I'm one of those weirdos that collects and reads the service manuals for classic Tektronix and HP test gear).

If totally open only means the software, aren't some x86 boxes there already, assuming you can replace the BIOS with coreboot?

If you can replace BIOS and firmware with free software, than you got a computer that runs only free software. "Totally open" is a vague term with no boundaries on its interpretation. I just specified what the post meant, not provided a definition for "open".

Maybe that's just me

Not just you. It would be awesome.

Interesting - I thought they were, but I can't find them either. Maybe it's only open-specification; I agree that the hardware isn't totally open without them.

There were also some issues with non-free Embedded Controller (the microcontroller that manages switches, power, clocks, and charging) firmware.

> Having a fast JS engine for MIPS could help these systems become more viable

Yeah, the lack of a modern, reasonable-speed JS engine for PowerPC is what eventually forced me to ditch my old PowerBook laptop, about a year or two earlier than I would've otherwise felt the need to do so (the next bottleneck would've been its inability to take more than 1.25 GB of RAM, but the web growing unusably slow happened first). It makes a huge difference in the viability of a platform, if you plan to use the modern internet on it, to have a current-gen JS engine, rather than an old-style fallback interpreter.

This will be interesting for the next generation of Loongson/Godson CPUs. Maybe the RMS approach of using Chinese Lemote Yeelong hardware driven by entirely open-source software will become more plausible for the rest of us.
I updated my very stale project wiki page with current project status, including the fact that we do now support the Loongson processors.

We've been testing this on the Lemote Yeeloong notebook, which has a Loongson 2F cpu. So far we have not done any specific optimizations for the Loongson, just making it work, by avoiding some of the mips32-specific opcodes that are not supported.

I'd be more interested in node.js (with V8) and mongodb on ARM.
Appears people have node working, judging from the comments on http://stackoverflow.com/questions/7483039/node-js-on-uclibc...

I havent tried it yet, but on todo list. Most stuff seems to be working, eg have redis working on arm.

The issue with javascript is that they do not support many architectures, other than interpreted only.

I'm curious to know, what modern devices are built on the MIPS architecture? (Our computer organization course includes the MIPS Instruction set, wanted to know how relevant or useful that might be.)