> From: beagleboard@googlegroups.com [mailto:beagleboard@googlegroups.com] On 
> Behalf Of Dennis Lee Bieber
> <robert.styles.forsyth-re5jqeeqqe8avxtiumw...@public.gmane.org> wrote:
> 
> >
> >For a screen based turnkey app, you can only display a splash screen for a
> >few seconds, before the user thinks it is broken, a bit longer for an
> >animated splash screen. Imagine a car radio, something has happen
> >immediately after power on (speaker click or LED indicator or screen
> >backlight), then less than half a second "Tuning..." before music appears.
> >
>       Probably not the best example. At best the "car radio" is running a
> microcontroller -- not a microcomputer -- with maybe an RTOS at best. The
> "radio application" is burned into the microcontroller flash memory, AND
> RUNS out of that flash memory. There is no setting up RAM (virtual memory
> mapping) as "RAM" in most microcontrollers is the general purpose register
> bank. There may be secondary flash or EEPROM used to store user settings
> (current volume/tone encoder position, stored station listing). A bigger
> unit may include a spinning disk or "user flash" for storing MP3s, but the
> core system doesn't have real "boot" phase.

I think the requirements concept is the issue.  Too many times in the past I've 
run into projects where the client makes a choice on the architecture and then 
we have to make the project fit.  Instead of a layout of the requirements and 
current/future capabilities as a design specification first.  Then look around 
at what is available to fill that.

Now if boot time is an issue then what amount of time is acceptable.  If it 
needs to be running in under 1 second then it doesn't matter how embedded the 
controller might be, a beagle or a pi running Linux is just not the tool for 
it.  Sort of...

Volume of systems sold is another criteria as well as development time and 
longevity.  The design decisions made if the device goes into a hard to access 
location will impact how it's designed and what's done with it. Etc.

I will state that waiting over two minutes for a Beagle to boot into Linux 
desktop mouse/keyboard/screen is unacceptable.  Windows with a 640K 8088 or a 
Pentium-33MHz  could create a graphical desktop way faster than a 1GHz Beagle 
with super fast SD card and 512MB of ram.  From an end user perspective 
anything longer than 10 to 15 seconds just means it's not done right.

So I'd like to pose the question back to the original poster of this thread 
karry.jais...@gmail.com.  Why are you using a Beagle if you need fast boot 
times?  What is it that the Beagle Green has that you need compared to 
something with an RTOS (Free RTOS for example).  Why Linux?

I've been on both sides of the fence.  I might have already posted the attached 
photo that shows a PiZeroW mounted onto a board that holds a PIC32, SuperCaps, 
RTC along with GPS module and SMS module.  Two antennas: one for GPS, one for 
SMS messaging.  And the connector to power and the vehicle CAN bus which 
requires instant on logging while the PiZero provides the file/networking 
infrastructure.  The Pi is the SPI master, the PIC32 is the SPI slave.  As long 
as the PIC has the RAM storage for all the CAN messages during the 18 second 
PiZeroW boot time all is good.  It wasn’t worth the time to port Microchip SD 
card file system support or all the network stuff needed for cell net access.  
We only built 20.

And with respect to University days I'll deny I ever took Fortran or Cobal 
courses.........

John




-- 
For more options, visit http://beagleboard.org/discuss
--- 
You received this message because you are subscribed to the Google Groups 
"BeagleBoard" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to beagleboard+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/beagleboard/00f601d750ed%244cafa650%24e60ef2f0%24%40autoartisans.com.

Reply via email to