So now RT is pretty regular everyday thing.
also people do not try to maximize the CPU times but to try giving more
in their CPU time.
it's like "I have 3M cycles just use them" if you are efficient you
indeed can use 2M cycles and it will do whatever the task is.
So there are Mission Critical Systems which require more CPU usage and
hence there shouldn't be any delay while providing a service.
On a Helicopter which is a more then RealTime system you will need about
3ms 100% guaranteed response time to a request response.
if it was done on 486 it's not a reason to make developers life harder
to be accurate and design their code in the same level of deficiency as
in xt.
Eliezer
On 06/22/2013 12:39 PM, babajaga wrote:> Not only medical or space. Just
a question of shortest response times. Which
> depends upon the HW generation. When doing RT-systems on old 16/32-bit
> systems this was in the range of (fractions of) microseconds. Because of
> instruction execution times. Which should be shorter nowadays.
> Besides, such things like "multi-threading, position independent code,
> reentrant code, spin locks, semaphores, manipulation of mapping registers
> etc." we had to do ourselves for RT-apps. In asssembler, of course.
So, all
> these functions are a very old "piece of cake", which finally found
its path
> into common programming practices :-)
>
>
>
> --
> View this message in context:
http://squid-web-proxy-cache.1019090.n4.nabble.com/Just-wondering-about-something-tp4660745p4660766.html
> Sent from the Squid - Users mailing list archive at Nabble.com.
>
