Advanced

Raspberry Pi GPU Driver Turns Out To Be Cr@p

Posted by Scott_Ruecker on · 13 comments
Phoronix; By Michael Larabel
Mail this story
Print this story

While it looked hopeful at first with today's announcement of a fully open-source graphics stack for the Broadcom VideoCore found in the popular Raspberry Pi development board, upon closer examination it's actually not that good.

Full Story

  Nav
» Read more about: Story Type: News Story

» Get new stories by e-mail: nightly or as they are published

« Return to the newswire homepage

  Related stories
» More like this

Comments (13)

Broadcom have a really cr@ppy support for Linux and Open Source. So it doesn't surprise me to discover that the hyped Open Source code release is nothing more than a shim.

At best this is half hearted, but seems more like a cynical exercise in PR.
So Broadcom put their whole user space core into a big "firmware blob" and then open sourced a loader for it. No one will likely be too impressed by this and it is good to see that the kernel folks are refusing to admit this into the kernel tree.

This looks like another way to get around gpl licensing and put your closed code into an otherwise open environment. Bahh..
Disappointing, to say the least. I could imagine a lot of applications for the Pi (or something Pi-like) using this chip -- but only if we can take advantage of it.
Wireless is becoming so dominant that RPi must have on-board wireless with open source driver to become really attractive for small devices. Having to add a wireless adopter/dongle is not a good enough option and the additional cost wouldn't be a problem.

Chi5t we really are twee now, removing obfuscated words from people's posts
@fettoosh --

I don't think this is about wireless. I think it's actually a Broadcom GPU.
@Dino,

I knew that but since it is about RPi resources, I thought it would be appropriate to wish for wireless on such a wonderful low cost SOC.

I am thinking about making use of it and I need wireless on it, I found this, unfortunately it uses up one scarce USB port.

Fettoosh, check out my first big project with the RPi:

http://mindplusplus.wordpress.com/2012/07/31/a-month-with-ra...

It's in the last section. Basically, turn a netbook (or laptop) and an RPi into mutual peripherals. Netbook is wireless proxy for RPi.
Thanks Gus3,

It is elaborate and will definately take a look. In the mean time, what do you think of Cubieboard? They say it will have wireless in sub board or next release. What is interesting to me is the GPIO, it has something like 96 pins.



@Fettoosh: How many of those GPIO pins can tie into IRQs? Any idea what the latency on those interrupts is likely to be? The device has I2C, so a few added PCF8574s can add a bunch of bare pins if you don't have to deal with them quickly, but IRQ-driving inputs are golden. Also, how are those GPIO pins structured -- 8051-like, with latch and driver, 8048-like with effectively a transparent latch, or blind-switched like an H8? Do they have 8051's Darlington-drive problem where, without a pullup, they lock low? (Yeah, yeah, RTFM -- where's TFM for this device?) Edit: n/m, found it: http://infocenter.arm.com/help/index.jsp?topic=/com.arm.doc....
Quoting: How many of those GPIO pins ... where's TFM for this device?


@cr,

good questions and I have no idea yet. I just noticed it yesterday and I am looking for input. The board is still not available except for some who contributed already. There are no details that I can see anywhere. But, you never know, it could be a big scam or it might be a real good option especially for the specs they claim to have below for only $50.

1G ARM cortex-A8 processor, NEON, VFPv3, 256KB L2 cache
Mali400, OpenGL ES GPU
512M/1GB DDR3 @480MHz
HDMI 1080p Output
10/100M Ethernet
4Gb Nand Flash
2 USB Host, 1 micro SD slot, 1 SATA, 1 ir
96 extend pin including I2C, SPI, RGB/LVDS, CSI/TS, FM-IN, ADC, CVBS, VGA, SPDIF-OUT, R-TP..
Running Android, Ubuntu and other Linux distributions


@cr,

Thanks. I guess I have a lot of reading to do. Opinions and ideas compared to RPi are very much appreciated.



Most of my questions are answered at the ARM site linked. I see no hardware reference, which stands to reason as ARM is a core-IP licensor, not a fab-house. You might be able to pull equivalent doc from licensees like TI and so gain an understanding of DC and transient characteristics of the port pins (such as the warning I noted in Intel docs about the Darlington problem). My embedded experience is pretty much with smaller (mainly 8-bit) MCUs and keyboard-controllers (and how they play with X86 BIOS), so when and if I can afford these toys we'll be learning in parallel. Just having the docs out there rather than buried in NDA is encouraging.

Post a comment

A free account is needed to post. Write your comment here first: your text is kept in this browser while you sign up or log in, and is waiting for you when you come back.

Your message:
|||||
10
Markdown: **bold**, *italic*, [text](https://link), - lists, > quotes, $x^2$ for math. Formatting help

Sign up and post (it is free)  or  Log in and post