[Debian] Switching tasking implementation?

[Debian] Switching tasking implementation?

[Jacob Sparre Andersen]

If I remember correctly, the version of GNAT distributed with Debian
supports more than one tasking implementation.

Do I remember correctly?  How do I select which implementation to use?

Greetings,

Jacob
-- 
"Two silk worms had a race. They ended up in a tie."

Re: Switching tasking implementation?

[Martin]

On Aug 26, 1:20 pm, Jacob Sparre Andersen <spa...@nbi.dk> wrote:
> If I remember correctly, the version of GNAT distributed with Debian
> supports more than one tasking implementation.
>
> Do I remember correctly?  How do I select which implementation to use?
>
> Greetings,
>
> Jacob
> --
> "Two silk worms had a race. They ended up in a tie."


Don't know about Debian but have you tried the configuration pragma
"Tasking_Policy"?

Cheers
-- Martin

Re: Switching tasking implementation?

[Martin]

On Aug 26, 1:25 pm, Martin <martin.do...@btopenworld.com> wrote:
> On Aug 26, 1:20 pm, Jacob Sparre Andersen <spa...@nbi.dk> wrote:
>
> > If I remember correctly, the version of GNAT distributed with Debian
> > supports more than one tasking implementation.
>
> > Do I remember correctly?  How do I select which implementation to use?
>
> > Greetings,
>
> > Jacob
> > --
> > "Two silk worms had a race. They ended up in a tie."
>
> Don't know about Debian but have you tried the configuration pragma
> "Tasking_Policy"?
>
> Cheers
> -- Martin

Sorry, I meant "pragma Task_Dispatching_Policy".

Cheers
-- Martin

Re: Switching tasking implementation?

[Jacob Sparre Andersen]

Martin wrote:

> Sorry, I meant "pragma Task_Dispatching_Policy".

I may not have been clear on what I want.  What I am interested in, is
to compare the performance of tasking based on OS threads with tasking
based on user-space threads.

Greetings,

Jacob
-- 
"[...] *transfer* a bit of salary from the person who writes
 a bug to the person that finds a bug..." -- Keith Ray

Re: Switching tasking implementation?

[Martin]

On Aug 26, 2:00 pm, Jacob Sparre Andersen <spa...@nbi.dk> wrote:
> Martin wrote:
> > Sorry, I meant "pragma Task_Dispatching_Policy".
>
> I may not have been clear on what I want.  What I am interested in, is
> to compare the performance of tasking based on OS threads with tasking
> based on user-space threads.
>
> Greetings,
>
> Jacob
> --
> "[...] *transfer* a bit of salary from the person who writes
>  a bug to the person that finds a bug..." -- Keith Ray

Ok - completely different thing then! And I've no idea, sorry... :-)

Cheers
-- Martin

Re: Switching tasking implementation?

[Ludovic Brenta]

Jacob Sparre Andersen wrote on comp.lang.ada:
> If I remember correctly, the version of GNAT distributed with Debian
> supports more than one tasking implementation.
>
> Do I remember correctly?  How do I select which implementation to use?

No, the two run-time libraries shipped in Debian allow you to choose
either zero-cost or setjump/longjump exception handling mechanisms.
The tasking model always uses the 1:1 model where each Ada task
corresponds to a kernel thread. I vaguely remember there used to be a
n:1 model where all Ada tasks would run in the same kernel thread,
doing scheduling in userspace. I'm not sure whether this model is
still available or not nowadays. Could you please look this up in the
GCC installation manual? This was a configure option IIRC.

--
Ludovic Brenta.

Re: Switching tasking implementation?

[Jacob Sparre Andersen]

Ludovic Brenta wrote:

> The tasking model always uses the 1:1 model where each Ada task
> corresponds to a kernel thread. I vaguely remember there used to be
> a n:1 model where all Ada tasks would run in the same kernel thread,
> doing scheduling in userspace. I'm not sure whether this model is
> still available or not nowadays. Could you please look this up in
> the GCC installation manual? This was a configure option IIRC.

Thanks.  It doesn't look like it is an option any more.

Greetings,

Jacob
-- 
"There is nothing worse than having only one drunk head."