[prev in list] [next in list] [prev in thread] [next in thread] 

List:       ucsimm
Subject:    [uCsimm] another note for RT-uClinux
From:       "=?euc-kr?B?sce8rrHZ?=" <kwonsk () mutech ! co ! kr>
Date:       2000-04-13 10:34:05
[Download RAW message or body]

Hi, all

i want to say that you need not feel unhappy after read my note
about 130 usec latency. Because it is only for timer task.
If you use rt_request_IRQ (like sampler.c) for hardware interrupt then
time for rtl_scheduler does not need (priority is resolved by hardware)
So for h/w interrupt latency might be near 50 usec.

And for timer task:
real world concern is not latency. If 130 usec latency is constant,
every timer event would be delayed 130 usec. therefore in view of
outside device there is no drift.  External device could not know that
every timer event is delayed 130 usec.  They just see constant interval
timer event.  In my guess and with experiment, 130 usec latency 
is stable, saying +/- 50 usec?

RT-uClinux could be said that (just my opinion)
1. time drift is bellow 50 usec
2. interrupt latency is about 50 usec.
3. lovely linux for non rt task.
and this might be comparable with commercial RTOS's.

But as you know,
it need more experiment. please do not trust what i said, trust yourself
and your own experimental result.

Thanks.
kwonsk.

This message resent by the ucsimm@uclinux.com list server http://www.uClinux.com/

[prev in list] [next in list] [prev in thread] [next in thread]