[Tinyos-help] Re: The relationship between radio sleep and CPU

[Cormac Duffy]

No the processor can be woken up by any external interrupt (see  
atmega128 manual, sure its the same for msp430)
therefor the cc2420 radio or similar will ususaly interrupt the  
processor when the packet arrives, dont know name of interupt of hand  
(see cc2420 manual).
This will wakeup cpu from power save and jump to interrupt. BUT  
certain power save mode such as sleep modes will take about 4 ms to  
wakeup, this might not be responsive enough. If this is the case use  
an idle power mode to reduce power consumption. In fact TinyOS  
already does this im sure automatically.
hope this helps
Cormac


On 5 Sep 2006, at 19:07, Yinying Yang wrote:

> Thanks.
> My understanding: The receipt of a message from radio can not wake  
> up the
> CPU from power-save state. But when CPU is in power-save mode,  
> radio can't
> forward a message. So if I would like to forward a received message  
> to other
> motes, I should set a timer to wake up the CPU periodically to  
> check whether
> there is a buffered message. If there is, forward it.
> Is it right?
>
> -----Original Message-----
> From: Cormac Duffy [mailto:c.duffy at cs.ucc.ie]
> Sent: Monday, September 04, 2006 4:00 AM
> To: yyyfau at gmail.com
> Cc: tinyos-help at Millennium.Berkeley.EDU
> Subject: Re: The relationship between radio sleep and CPU
>
>
> No the radio can operate independently, at least on most platforms,
> therefor its usually important to have a coordinated duty cycle,  
> between
> the processor
> and the radio.
>
> Also it is possible to have the processor in a low-power (sleep or  
> idle
> state) and get the radio to interrupt, (there by waking up the
> processor,) when it senses packets on the radio channel, ie with  
> carrier
> sense. Obviously the radio can buffer a packet until the processor  
> wakes
> up and the
> processor can then process packet. I am unsure if in tinyos the radio
> drivers use a boradcast address and determine the the receipient after
> packet downloaded from radio, OR weather the radio determines who  
> is the
> receipent of the packet. Obviously the later would be better,  
> because it
> means the radio wouldnt wake up the processor to see who the  
> packets for.
>
> I don't specialize in MAC protocols.
> Hope that was a help
> Cormac
>
> On 3 Sep 2006, at 20:02, Yinying Yang wrote:
>
> Message: 1
> Date: Sat, 2 Sep 2006 21:05:33 -0400
> From: "Yinying Yang" <yyyfau at gmail.com>
> Subject: [Tinyos-help] The relationship between radio sleep and CPU
> 	sleep
> To: tinyos-help at Millennium.Berkeley.EDU
> Message-ID:
> 	<90bc03d20609021805v5e2e4b24k15c674c172d4ffe6 at mail.gmail.com>
> Content-Type: text/plain; charset="iso-8859-1"
>
> HI,
>
> When we put the mote into CPU sleep (power save) mode, would the  
> radio be in
> sleep mode too?
> Or it is possible when we put the mote into CPU sleep mode, while  
> the mote
> can still send and receive messages?
>
> Thanks.
>
>
> -- 
> Cormac Duffy     B.Sc, M.Sc
> Computer Science Dept.
> University College Cork,
> College Rd.,
> Cork,
> Ireland.
>
> email:            c.duffy at cs.ucc.ie
> web:		     www.cs.ucc.ie/~cd5
> telephone:     0872039750
>
>

___________________________________________________

Cormac Duffy M.Sc B.Sc

Computer Science Dept.

University College Cork

College rd.

Cork

Ireland



email:             c.duffy at cs.ucc.ie

web:               www.cs.ucc.ie/~cd5

mobile:          +353-87-2039750


-------------- next part --------------
An HTML attachment was scrubbed...
URL: http://mail.millennium.berkeley.edu/pipermail/tinyos-help/attachments/20060905/4772100a/attachment.htm