[Ace-users] measurement with ACE_High_Res_Timer

[perrone]

On 2 dez, 23:27, "Sandro Andrade" <sandro.andr... at gmail.com> wrote:
> Hello,
>
> I suppose Ricardo is looking for a non-preemptive execution of his "block of
> code" in order to estimate execution times. Something like using a FIFO
> scheduling with the highest priority.
>
> Ricardo ?
>
> Thanks,
> Sandro Andrade
>
> On Dec 2, 2007 11:47 PM, Douglas C. Schmidt <schm... at dre.vanderbilt.edu>
> wrote:
>
>
>
> > Hi Ricardo,
>
> >   Thanks for using the PRF.
>
> > > CIAO version 0.6.10
> > > TAO version 1.6.10
> > > ACE VERSION: 5.6.1
>
> > > HOST MACHINE and OPERATING SYSTEM:
> > >       Centrino Duo Core, Ubuntu 7.0.4-Feisty, kernel 2.6.20-16-generic
>
> > > COMPILER NAME AND VERSION (AND PATCHLEVEL):
> > >       gcc (GCC) 4.1.2
>
> > > THE $ACE_ROOT/ace/config.h FILE:
> > >       #include "ace/config-linux.h"
>
> > > THE $ACE_ROOT/include/makeinclude/platform_macros.GNU FILE:
> > >       no_hidden_visibility=1
> > >       include
> > >  /usr/local/ACE_wrappers/include/makeinclude/platform_linux.GNU
>
> > > CONTENTS OF $ACE_ROOT/bin/MakeProjectCreator/config/default.features:
> > >     [there is no default.features file, but the gnuace and global
> > >  features]
> > >         qt=1
> > >         fl=1
> > >         tk=1
> > >         xt=1
> > >         cidl=1
> > >         sctp=1
> > >         zlib=1
>
> > > DOES THE PROBLEM AFFECT:
> > >      COMPILATION? no
>
> > > SYNOPSIS:
>
> > > DESCRIPTION:
>
> > >In ACE_High_Res_Timer class i don't see anything that prevents an
> > external interference between start and stop methods. I  need to measure a
> > block of code like an atomic operation.
>
> > What do you mean by "atomic operation"?
>
> > > How can i do this? The sample code below depicts a measurement process:
>
> > >ACE_High_Res_Timer::start ();
> > >block of code;
> > >ACE_High_Res_Timer::stop ();
>
> > >is The ACE_Sig_Guard class  capable to guarantee this level of isolation
> > or is there another better solution?
>
> > It's not really clear what you are trying to be atomic wrt, e.g.,
> > signals, thread context switches, etc., so can you please clarify this
> > first?
>
> > Thanks,
>
> >        Doug
> > --
> > Dr. Douglas C. Schmidt                       Professor and Associate Chair
> > Electrical Engineering and Computer Science  TEL: (615) 343-8197
> > Vanderbilt University                        WEB:
> >www.dre.vanderbilt.edu/~schmidt<http://www.dre.vanderbilt.edu/%7Eschmidt>
> > Nashville, TN 37203                          NET: d.schm... at vanderbilt.edu
>
> > _______________________________________________
> > ace-users mailing list
> > ace-us... at mail.cse.wustl.edu
> >http://mail.cse.wustl.edu/mailman/listinfo/ace-users
>
> --
>
> Sandro Santos Andrade
> --------------------------------------------------------
> Distributed Systems Laboratory (LaSiD)
> Computer Science Department (DCC)
> Federal University of Bahia
> Brazil

Hi Sandro,

you're right.... but FIFO scheduling on no-preemptive linux kernel can
be  temporarily interrupted by kernel's I/O threads, right?

Thanks,
Ricardo