Anyone Wanting to Try Rasberry Pi ?

No it's not a sex act (I hope)... it's essentially an ARM processor on a board (Like Arduino the board) except it comes with USB, HDMI, a GPU etc.. basically a phone without a screen but with real world connectors you see on computers, you plug your keyboard and mouse into it...

(without going into shock over questions like this... http://www.raspberrypi.org/forum/general-discussion/worthwhile-me-buying-a-pi-want-to-learn-a-language-for-running-calculations#p64804 )

What's peoples opinions? flop or possible success?.. of course $25 for a 700mhz ARM processor and GPU i'm going to eventually buy one out of curiosity, but for those who've not heard of it, have a look at this, they clam that they're pushing this at kids to get them interested in programming again, i don't see how personally, I just want to write one for an extremely low cost low powered web server with 256 meg of ram, i can't quite imagine the Atmega putting up much of a fight.. of course this is not a replacement, the beauty of any Arduino board is the header pins to interface with the real world... so forget it, nobody's taking my Arduino's away from me NEVER!!!!....

http://www.raspberrypi.org

Maybe some kind of hack on the Rasberry Pi to make it interface with Arduino? my electronics side of me says Rasberry what.... my geek side says, whoooooo $25 gets you a 700mhz CPU and GPU and 256 meg or ram which runs Linux direct from SD card... so to anyone who thinks this might be a threat to Arduino's future forget it, unless we see header pins to interface with the real world, the rasberry pi will just be another event in the computer history calendar, let's face it an UNO board as of now will be just as suitable to do stuff on now or in 10 years time, you don't need to go faster (we don't run an OS on it...) with a phone or this rasberry pi, newer software comes along and the device will fail... if a new sensor comes along and we can't get the data off... that sensor fails :)

The Raspberry Pi is surely going to affect how people use Arduinos. Just like the Arduino it appeals to the hobbyist on price and ease of use, but suddenly we can opt for a whole computer complete with operating system and video output. I don't think Arduino can compete on price, but it does have certain strengths:

  • Plenty of I/O ports
  • Can run at very low power
  • Better suited to real-time tasks

I hope the RPi will stimulate some evolution among the existing Arduino species. I think we'll see lots of Arduino-based I/O expansion boards for the RPi, for a start. Perhaps we will also see newer Arduinos that play to their strengths, such as remote sensing or mobile applications, with support for Lithium Polymer batteries, solar panels, or low-cost rf communication.

In any case, it's a great time to be an electronics and computing hobbyist.

I'm waiting for the initial riots to die down.

Although, frankly I don't understand a lot of the excitement. You've been able to get similar systems at not-too-much-higher prices for quite a while (in the form of linux-based appliances: routers, chumbies, pogo-plugs, etc.) And if size isn't so much of an issue, you can frequently get a PC-class machine for free from someone who is upgrading. I've probably got a half-dozen things lying around that could fill a "I want to learn python programming under linux" urge (not counting the actual computers that could run linux under VirtualBox or the equiv.)

I'm kind of annoyed at the choice of a broadcom chip which you aren't actually allowed to find out everything about without doing a deal with broadcom and signing NDAs (and probably paying a fortune). Seems counter to the sort of thing the project is supposed to be about (but no doubt it will be fun for people who like to reverse engineer things :-).

Here, here.

Yes, it's interesting and I'll probably buy one. (After things settle a little.) But, what about the Beagle Bone? It's a bit pricey in comparison, and I don't know whether it does 1080p decoding, but otherwise, it's an Arduino running on a 700MHz ARM processor. GPIO is already available on the BB. It's an option for the RP.

So, if the AVRs were going to be threatened by a core clock speed makeover, that ship would have sailed by now. Fact is, there's plenty you can do at 16MHz. There's no need for near-GHz speed for many of these projects. There's plenty of room for low-power processors. Right tool for the job and all that.

I would also like to strongly second the discontent with proprietary chipsets, though I don't know that there's any option out there for hardware-accelerated 3D and video decoding that isn't going to be swamped in patents and NDAs. That's what we, as a society, get for allowing the fancy HDMI, Dolby, and MPEG logos to sway us away from perfectly good open and patent-free alternatives. Now, if you want to play commercial content, you have to be licensed to decode those formats. That costs money. To ensure those fees get paid, the libraries aren't available for free. (Part of the Pi's cost is licensing of MPEG4/h.264 codecs. No MPEG2 yet.) And since there's no point in restricting release of binaries if the source is available, you don't get that either. Which means you're stuck with the drivers and kernels made available to you. Happy tinkering! Far cry from the BBC micro in that sense.

Sorry. Touchy subject for me. :cold_sweat:

Claghorn: I'm kind of annoyed at the choice of a broadcom chip which you aren't actually allowed to find out everything about without doing a deal with broadcom and signing NDAs (and probably paying a fortune). Seems counter to the sort of thing the project is supposed to be about (but no doubt it will be fun for people who like to reverse engineer things :-).

Ok, what other ARM+GPU could they have used that has an open GPU datasheet?

I'll be having one when you can buy one straight off, It'll be going at the other end of the USB lead to the arduino and doing things the arduino doesn't do well if at all. (Anything involving more than 2K of ram, ethernet (lets not pretend that the ethernet shield is actually usable), an RTC (OK it doesn't have one of its own but its pretty straightforward to use NTP and the internet and keep an accurate clock), plotting nice graphs from the data the arduino throws at it, logging stuff to an SD card or HDD ). Sort of a 32 bit all purpose shield......

Hey, the Arduino can do NTP, and I managed to get the web server up to an almost-reasonable speed. It can even plot graphs. For a while I had one plotting weather graphs and web-serving the results.

But really an Arduino wasn't the best solution for a home web-server. And neither was my old, unloved, noisy and power-hungry PC. This is where the RPi will be perfect: small and economical, but powerful enough to run data analysis software, web-servers, and plenty of other things.

The argument about learning Python et al. is I think only half the story. After all, you can install a Linux virtual machine on almost any modern PC and hack away to your heart's content. The point of the RPi (IMHO) is that once you've made it do something, it's cheap enough to install it somewhere and leave it doing that thing: whether it's flashing lights, watering the garden, web-serving data from your home weather station, turning your TV into a digital picture frame, or a zillion other things. In that sense it's similar to the Arduino, but the possibilities it opens up are somewhat different.

Now the developers of the Beaglebone, Chumby and other boards can justifiably feel a bit peeved about all the attention RaspberryPi is getting. Many are arguably more powerful than the RPi. But the RPi people have got a lot of things right - in particular the connectivity, and the price. I think they've reached a certain threshold of appeal, and that's the reason for their success.

I'm kind of annoyed at the choice of a broadcom chip

This is because the guy who's baby this is works for broadcom.

Ok, what other ARM+GPU could they have used that has an open GPU datasheet?

Well the ST one for starters.

Its got a lot of publicity because the Beeb picked up on it and its parallels to the BBC micro 30 years ago. I can't see GM being over impressed given his experience with said 30 year old computer.... ;)

When the final crush dies off we have a "need" for 3 here as my first order. Even as a basic PVR and internet TV upgrade for old TVs out of the blocks RPi has a lot of potential. And as others have said, it's an awful lot of capability at very near the Arduino price point. I also have some old PIII hardware running Linux about the place currently for light duties and tinkering that can hopefully be retired and reduce our power consumption (and the heat in this room!).

Beaglebone and others are interesting, but the price point of RPi will I'm certain make a lot of people try it out and that can only be good for Arduino (at least in the short term) and hobby electronics overall. In the longer view, if the Raspberry Pi foundation can execute, and if the platform can live up to most of the initial hype, and once the RPi community grows and the ecosystem matures, Arduino could well become the remote sensing/hardware interface card of choice for RPi. It will certainly shake things up.

As for the choice of Broadcom, they used what they were familiar with and a company they had close ties to. That sounds sensible to me. Remember too, that we're not the ultimate target demographic.

I hope they can pull it off, Geoff

To be fair the first Raspberry Pi prototypes didn’t use Broadcom chips. They used… (wait for it) …an ATmega! Yes, the first RPi was just a big Arduino. How about that.

Well the ST one for starters.

ST has an ARM with a GPU (Graphics Processing Unit)? Note that the ARM11 used in RPi is quite a different beast than the Cortex M0/M3/M4 parts that are usually aimed at microcontroller replacement. ARM11s are aimed at unix-running (and similar OSes) "appliances" like cell phones, cable TV boxes, Digital Video Recorders/players. and stuff like that. ARM11 is a little obsolete (the Cortex A8, as used in beagleBoard, is its nominal replacement, I think.) But that also means that it's "mature" and companies like Broadcom have those lovely System-On-a-Chip (SOC) chips with all sorts of large-system peripherals (GPU, MMU, DRAM controller, USB host, etc.)

Grumpy_Mike:

Ok, what other ARM+GPU could they have used that has an open GPU datasheet?

Well the ST one for starters.

I can't find an ST ARM with a GPU on ST's site http://www.st.com/internet/mcu/class/1734.jsp.

As far as I know, all of the ARMs with GPUs are locked down and under NDA. It's not a good situation but it's where we're at at the moment.

ARM does have a standard GPU, don't they? Neon?

I think the bit about RP being used for educational purposes (vs. a PC or VM of Linux) is that you can outfit a classroom with RPs, KBs, mice, and cheap TVs for a bit less money than a room full of computers. Especially when you consider maintenance costs.

It'll also work for the DIY crowd, because it's a fairly complete mini-PC platform -- even the Linux distro is usable out of the box, with the provided images. That removes quite a bit of the uncertainty for those less than familiar with hardware and Linux to get to the point where they can follow online walkthroughs with a minimum of effort. The exception being DVR and IPTV use. Without MPEG2 licenses (read: support), it's useless. (Anyone here dealing strictly with VC-1/AVC yet? Not me.) And I don't think a 700MHz ARM has any hope of handling it in software.

As a hacker's toy, this is not the ideal platform. Sorry, but unless you can get down to the silicon level, it's still a black box. That rubs against the grain for me, especially when it's touted as the ultimate experimenters kit. Look, either I'm allowed to take it apart or I'm not. This halfsees stuff doesn't cut it. Sure, you can peek under the hood, but you're confined to running compatible kernels because someone doesn't want to give away their trade secrets. Fine. Then it is what it is, and (IMO) what it is, isn't for hackers.

But, I think dhunt hit it on the head... What alternatives are there for full-featured GPUs? Technology at this level is just always encumbered by patents. Frustrating how much further a community could take a technology if there weren't always legal roadblocks in the way.

Ill tell you now, typing 80 column text or anything higher resolution on a cheap composite TV will drive you away in about 20 seconds, so replace cheap TV with couple hundred dollar HDMI enabled HD TV

Cause I myself went though oh about 5 composite TV's over the last 2 years before finding one that actually useable, and its been out of production for a decade

Thats my biggest beef, no VGA or even DVI

DVI to HDMI (or visa versa) adapter ($5 bucks, problem solved) .. or even HDMI to VGA (if they're around)

HDMI to DVI is cheap and easy since they are both digital. HDMI to VGA is much more difficult and expensive since VGA is analogue. I can see their point that 'VGA is end of life' but the trouble is VGA has been 'end of life' for over a decade, it just refuses to lie down and die. It would make the Pi much more expensive to put VGA on it, and previous posters are right. the analogue composite video option is basically c**p. I have an old 4:3 aspect monitor sat under my bench that has DVI, so that'll do me. Ultimately it'll be controlled by SSH, so the display won't matter.

http://www.ebay.co.uk/itm/DVI-MALE-HDMI-FEMALE-ADAPTER-CONNECTOR-CONVERTER-/270890136395?pt=UK_Computing_CablesConnectors_RL&hash=item3f124f734b#ht_3051wt_1141

VGA is much more difficult and expensive since VGA is analogue.

so is composite, so they are taking a RGB space and squashing it down to a color burst and sync's all on one wire, seems more trouble unless the chip just wont do a vga signal

I can see their point that 'VGA is end of life'

yea and composite has been on the chopping block since the early 90's with SVIDEO and SCART, it just seems to be a very odd choice for a computer system other than media playback on a SDTV, and even then it could be better with the above outputs

shrug