TRON for AdaOS

TRON for AdaOS

[Marius Amado Alves]

Nick Roberts, David Botton, and all:

Have you considered the TRON Specifications for AdaOS?

Just an idea that may or may not speed up AdaOS construction, and at 
the same time tap into a somewhat different but striving (at least in 
Japan) market.

The TRON Specifications are a set of open specifications of real-time 
operating systems. Initially developed at Sakamura Labs I think. 
Personal Media Corp. (also Japanese) now offers the T-Engine Kit at 
circa 1500 euros. The T-Engine Kit is a TRON-based combination of 
hardware and software for system development.

Re: TRON for AdaOS

[Marin David Condic]

Would you happen to have any links or textbook references?

I like the idea of an Ada RTOS - although others have been built. (Might 
be better to start from something like RTEMS & build on it?) There are 
also various realtime Linux implementations one might learn from. At 
least one implementation I have heard of runs a single realtime process 
and then coughs up any leftover CPU time to run the rest of Linux. I 
hear it works pretty well and provides a basis for an OS that does 
something Windows doesn't do. You've got to think of why anyone would 
want an Ada OS besides the fact that it is implemented in Ada.

MDC


Marius Amado Alves wrote:

> Nick Roberts, David Botton, and all:
> 
> Have you considered the TRON Specifications for AdaOS?
> 
> Just an idea that may or may not speed up AdaOS construction, and at the 
> same time tap into a somewhat different but striving (at least in Japan) 
> market.
> 
> The TRON Specifications are a set of open specifications of real-time 
> operating systems. Initially developed at Sakamura Labs I think. 
> Personal Media Corp. (also Japanese) now offers the T-Engine Kit at 
> circa 1500 euros. The T-Engine Kit is a TRON-based combination of 
> hardware and software for system development.
> 

-- 
======================================================================
Marin David Condic
I work for: http://www.belcan.com/
My project is: http://www.jsf.mil/NSFrames.htm

Send Replies To: m   o   d   c @ a   m   o   g
                    c   n   i       c   .   r

     "'Shut up,' he explained."

         --  Ring Lardner
======================================================================

Re: TRON for AdaOS

[Marius Amado Alves]

Primary TRON sites (in my understanding):

    http://www.tron.org
    http://www.sakamura-lab.org
    http://www.personal-media.co.jp

If your Japanese is, like mine, a bit rusty :-) check these items:

    http://tronweb.super-nova.co.jp
    http://www.linuxinsider.com/story/31855.html

Regarding real-time power and RTLinux, this excerpt from the latter 
item deserves spotlighting:

"RTLinux switches tasks in milliseconds, while ITRON switches tasks in 
microseconds," he said. "RTLinux' footprint is measured in megabytes; 
ITRON is measured in kilobytes."

Re: TRON for AdaOS

[Ole-Hjalmar Kristensen]

>>>>> "MAA" == Marius Amado Alves <amado.alves@netcabo.pt> writes:

    MAA> Primary TRON sites (in my understanding):
    MAA>     http://www.tron.org
    MAA>     http://www.sakamura-lab.org
    MAA>     http://www.personal-media.co.jp

    MAA> If your Japanese is, like mine, a bit rusty :-) check these items:

    MAA>     http://tronweb.super-nova.co.jp
    MAA>     http://www.linuxinsider.com/story/31855.html

    MAA> Regarding real-time  power and RTLinux,  this excerpt from  the latter
    MAA> item deserves spotlighting:


    MAA> "RTLinux switches tasks in milliseconds, while ITRON switches tasks in
    MAA> microseconds," he said. "RTLinux'  footprint is measured in megabytes;
    MAA> ITRON is measured in kilobytes."

I would be very surprised if "RTLinux switches tasks in milliseconds"
actually is true. On my system, (a vanilla Linux admittedly :
2.4.9-e.34smp #1 SMP ), task switching takes about 4.5 microseconds.

-- 
   C++: The power, elegance and simplicity of a hand grenade.

Re: TRON for AdaOS

[Vinzent 'Gadget' Hoefler]

Ole-Hjalmar Kristensen wrote:

> I would be very surprised if "RTLinux switches tasks in milliseconds"
> actually is true. On my system, (a vanilla Linux admittedly :
> 2.4.9-e.34smp #1 SMP ), task switching takes about 4.5 microseconds.

And that timing is guaranteed?


Vinzent.

-- 
worst case: The wrong assumption there actually is one.

Re: TRON for AdaOS

[Ole-Hjalmar Kristensen]

On my system, it's surely not guaranteed. But it's perfectly adequate
for building soft real time systems.
My point is that the task switching overhead itself is very low on
Linux. I did not say anything about the granularity of the time slice.
My test program also includes the overhead in accessing a shared
protected object, btw.

>>>>> "V'H" == Vinzent 'Gadget' Hoefler <nntp-2005-02@t-domaingrabbing.de> writes:

    V'H> Ole-Hjalmar Kristensen wrote:
    >> I would be very surprised if "RTLinux switches tasks in milliseconds"
    >> actually is true. On my system, (a vanilla Linux admittedly :
    >> 2.4.9-e.34smp #1 SMP ), task switching takes about 4.5 microseconds.

    V'H> And that timing is guaranteed?


    V'H> Vinzent.

    V'H> -- 
    V'H> worst case: The wrong assumption there actually is one.

-- 
   C++: The power, elegance and simplicity of a hand grenade.

Re: TRON for AdaOS

[Vinzent 'Gadget' Hoefler]

Ole-Hjalmar Kristensen wrote:

> My point is that the task switching overhead itself is very low on
> Linux.

Yes. But such things can change when you apply real-time requirements.
Sometimes it is very surprising to see the difference between an
average-case optimization and a worst-case optimization. ;-)

Well, I searched for it and the context switching times simply seem to
depend on the method used.

<URI:http://trecom.upv.es/articles/NTCC150-4.PDF> states:

|The communication method between the scheduler and the Linux process
|can be done by means of the RT-Linux standard RT-fifos, although a
|shared memory mechanism has also been implemented in order to improve
|efficiency. This method allows the scheduler to improve context
|switches by three milliseconds below the accomplished with the FIFO
|communication.

So indeed here they talking about milliseconds, but with the shared
memory approach this seems to change considerably:

|Some measurements about context switching times in RT-Linux have been
|made in the execution of the previous program. [...] In the case of
|interrupts, this time is very variable, but it can be up to 20
|microseconds in the worst case. [...] In the case of system calls, the
|context switching time is much lower: 8 µs. The scheduler requires
|about 4 microseconds.

So that doesn't sound too bad. ;)


Vinzent.

-- 
worst case: The wrong assumption there actually is one.

Re: TRON for AdaOS

[Marin David Condic]

See: http://www.linuxdevices.com/articles/AT8916798258.html

The RTLinux Manifesto seems to suggest performance is pretty good. I 
know of it being used in real world hard realtime applications, so it 
can't be all that bad. It may not be the best or only answer so Tron may 
have something going for it too.

Of course the real problem is that there is no truly active and 
progressing effort to build a usable Ada OS, so it all gets kind of 
academic as to whether or not RTLinux or Tron or anything else can help 
out in some regard.

MDC

Ole-Hjalmar Kristensen wrote:

> 
> I would be very surprised if "RTLinux switches tasks in milliseconds"
> actually is true. On my system, (a vanilla Linux admittedly :
> 2.4.9-e.34smp #1 SMP ), task switching takes about 4.5 microseconds.
> 

-- 
======================================================================
Marin David Condic
I work for: http://www.belcan.com/
My project is: http://www.jsf.mil/NSFrames.htm

Send Replies To: m   o   d   c @ a   m   o   g
                    c   n   i       c   .   r

     "'Shut up,' he explained."

         --  Ring Lardner
======================================================================