[En-Nut-Discussion] any real case of 4 UART ports?

[Harald Kipp]

Danny,

it's not a question of the number of UARTs but of
expected throughput, e.g. baudrate.

115 kBaud will produce 11,500 cps. On Ethernut 1.3 and 2.x
this results in 1,282 cycles available for each character.
Thus a lot of code can be processed for each character.

In worst case we get 4 times more incoming characters and
the same amount to send. Then 160 cycles are left for each
character, which is not much, but still practicable.

Fortunately the 1655x chips got internal buffers of at
least 16 characters. If your driver reads all buffered
characters within the interrupt routine instead of
getting interrupted on each single character, there will
be plenty of time. Furthermore you might consider to
run the transmitter in polling mode and avoid transmit
interrupt overhead.

On the other hand, what will an Ethernut board do with
46,000 characters per second? How is this relatively
large amount of data processed? What I mean is, you
must consider two parts: The UART I/O and the other
side, whatever that might be.

Harald

At 17:08 17.09.2004 +0800, you wrote:
>I now extend ethernut's UART up to 4 by using a ST16554 controller,
>but I wonder the Mega128 can simultaneously process data from four UARTs, 
>so does anyone have real case to proved that?
>Best Regard
>Danny Lu
>Pou-Yuen Tech. Inc.
>TEL: 886-4-23580345 #310
>FAX: 886-4-23580333
><mailto:danny.lu at pouyuen.com.tw>danny.lu at pouyuen.com.tw
>_______________________________________________
>En-Nut-Discussion mailing list
>En-Nut-Discussion at egnite.de
>http://www.egnite.de/mailman/listinfo.cgi/en-nut-discussion