BeagleBoard-X15

Another possible target for RISC OS.
The BeagleBoard-X15 has just been announced. Due to be available February next year.
See here.

2x 1.5GHz Cortex-A15
2GB ram
4GB eMMC
eSATA port
2xGigBit Ethernet
3xUSB 3.0 ports.
1xHDMI

The TRM for the AM572x is available here.

The TRM for the AM572x is available here

Just starting at page 349 (i.e. skipping the contents/index pages…)
Hardware floating point¹ x 2, Gb ethernet x 2.
“– Streaming video up to full high definition (Full-HD) (1920×1080p, 60 fps)”²

So who is leaping forward to do the HAL work for that?
Ah…

¹ Audio processing is referenced
² Should keep Rick M happy

“– Streaming video up to full high definition (Full-HD) (1920×1080p, 60 fps)”2

Mmm, any Makoto Shinkai production on that and a big monitor. Mmmmmmm! ;-)

So you’re going to do the HAL then? :)

Hmm, the memory map looks kinda similar to the OMAP5. So it might be possible to extend the OMAP5 HAL to work with it, or at least use it as a jumping-off point for a new HAL. The obvious main benefit over OMAP5 being that the ethernet is in hardware, not USB.

So it might be possible to extend the OMAP5 HAL to work with it, or at least use it as a jumping-off point for a new HAL.

Oh, distraction? Or were you thinking of coaching someone?

Chris why do you keep teasing us like this. It would be interesting to find out if a hard disc connected via sata 3G is faster than USB3.0 – well I can dream of getting disc storage up toward 500MB/s on RISC OS can’t I? I’d settle for maxing out USB2 at the moment.

I liked the bit in the article where the writer would be surprised if it cost more than 100$ – it’s not April 1st is it?

So it might be possible to extend the OMAP5 HAL to work with it, or at least use it as a jumping-off point for a new HAL.

Oh, distraction? Or were you thinking of coaching someone?

Neither. I already have too many distractions, including coaching people!

I liked the bit in the article where the writer would be surprised if it cost more than 100$

I liked the bit where he said “it would be surprising if the X15 is priced at over $100”¹ and in the very next paragraph, discusses a potential move of the BeagleBone Black to a Broadcom chip because “TI is not happy to sell to them a low cost processor”. The two points of view don’t add up.

Would a move to a BCM SoC be good? On the one hand, there’s the No – the documentation. TI’s level of documentation is quite lovely and BCM has yet to see the light. On the other hand, there’s the Yes – because if it is anything like the Pi’s SoC, the video will just work. I would use the Beagle more if it wasn’t a battle to get anything to appear on the screen (seriously – S-Video is a better option) and even then, the more powerful OMAP3 struggles to deal with my 1280×1024 display. I tend to need to run it at 1024×768. The OMAP3’s inability to deal with large displays definitely points to its intended use in small mobile devices; my first Android phone, the Motorola DEFY has an OMAP3630 inside and the resolution is 480×854 with 480p video recording. It cannot handle playing back 720p HD.

I also like the assertion that “The original BeagleBoard pretty much single-handedly started the trend of open-spec Linux SBCs”.
Yes, the Beagle was one of the first, but its price meant that it was something for developers and the better off. This was a trend that boards followed until the RaspberryPi came along and put a usable SBC into everybody’s reach (even children). Not only that, but one powerful enough that a few clip together bits of plastic and a twiddle or two and you have enough to be a competent media player that you can just tape to the back of the TV.

That said, the idea of an SBC existed before the Beagle. Something I might credit with being one of the first was my video recorder. Based upon a TI chip, it ran a cut-down Debian and came with stuff like telnet baked in as standard. So it was not difficult to port the Brandy interpreter² to allow simple BASIC programs to run on it. If the datasheets were fully available back then, I reckon it could have been taken further – both as a video recorder and as a small computer board. With on-board LAN, 64GB RAM, and CF/SD/USB, it might have made a good lightweight web server. Okay, it wasn’t designed to be an SBC³, but it was designed to be hackable.

¹ Translation for British readers: $100 will be a little over £100, strangely enough, even though Google reckons it should be about £63. Must be priced in petrodollars or Canadian dollars or something. Yeah, that’s what it’ll be…

² Doesn’t quite get line endings correct all the time. I think this is a CRLF→LF issue.

³ Trying to attach a keyboard/mouse to the USB port is fail – there’s no support in the firmware for that. Best done via telnet console…

That said, the idea of an SBC existed before the Beagle.

Go back to the BBC micro era. Or before. Things are just smaller now.

That said, the idea of an SBC existed before the Beagle. Something I might credit with being one of the first was my video recorder.

There’s also the LART which looks like it would have predated your PVR by a few years. Slightly less user-friendly IO, but it fully embraced the open hardware concept.

And speaking of PVRs (or better: STBs), the first “hackable” platform was surely the d-Box 2 (first hardware sold in 1999) which, because the original BetaNova software was so horrible, became the first hardware to run the TuxBox project. I think it was in 2000 or 2001. PowerPC 66 MHz, 64 MB RAM, half duplex 10 MBit ethernet. And a 56k modem.

There’s also the LART which looks like it would have predated your PVR by a few years.

Brilliant! Thanks. I knew there was a Linux board that predated the Beagle by a good long while, but I couldn’t remember the name.

Go back to the BBC micro era.

It was pretty much all single-board hardware back then. The Master Compact came along and bucked a trend, although to be fair the box at the back didn’t do a lot. ;-)

It was pretty much all single-board hardware back then.

The thing is it was before half of those Linux guys¹ were old enough to require nappies. ;)

¹ Shortening the sentence at this point is an option.

Nota: a low cost single core version of the BeagleBoard-X15 is announced… If price is good, Beagleboard-X15 could be the best option for a low cost and high performance RISC OS computer (after the Raspberry Pi 2, of course).

Now listed on the digi-key website

The official page to watch for BeagleBoard-X15 progress is http://www.elinux.org/Beagleboard:BeagleBoard-X15.

The German reading minority can follow my RISC OS blog, latest BeagleBoard-X15 article is http://riscosblog.huber-net.de/2015/07/neuigkeiten-zum-beagleboard-x15/

Price seems to be around 200 Euro, so a bit more expensive than some optimists expected. But a lot cheaper than the ISEE IGEPv5.

I did expect this price. The point is that they suggest that a future mono-core version could be available later at a lower price.

Interesting that the X15 uses the same chip as Titanium

Even more interesting is that whilst it is not available yet Mouser have it priced at 140ukp ex vat

Beagleboard-X15 is now available (without FCC and CE testing) here at a mere £173.99. Not sure how much more the properly tested boards will be when they become available in a month or so.

Does the work on Titanium mean that a RISC OS port is imminent?

Not sure how much more the properly tested boards will be when they become available in a month or so.

$239 (i.e. about £200 at the current rate of exchange). The ‘month or so’ is now said to be ‘start shipping on 23 September 2016’. This was referred to as ‘February next year’ in November 2014 (see original post on this thread).

As it appears so similar to the Titanium, is a RISC OS port a trivial step now?