[Developers] Proposed Cactus Timer API Completion

Steve White swhite at aei.mpg.de
Fri Aug 6 03:36:59 CDT 2004


On Thu, 5 Aug 2004, Gabrielle Allen wrote:
> remember that a lot of additional things can be implemented in a Timer
> thorn to keep the Flesh slim ...
I see that, and I'm already tempted to write a couple of clocks myself.
MPI timers spring to mind...I know people are using them.

BTW, I also appreciate the move to present a unified interface for timing.  
The underlying 'gettimeofday' and 'getrinfo' are themselves a mess to use.

> The idea was to have a thorn which would also provide the information  
> in different customizable ways, including standard timers for different  
> IO, reductions, communications etc, and including writing it to file,  
> displaying it on the HTTPD interface, on the Cactus portal,
Isn't this what the PerfReports thorn is for?  I think all of this is
great, but I'm not sure what it has to do with the Timer interface.
> and sending notifications to people when e.g. their MFlops dropped to
> an embarrassingly low value.
This would be the job of an Alarm or Alert.  Such a thing might refer to a
Timer, but (once again) a Timer shouldn't be concerned with MFlops.  
Maybe a thorn with a name like SystemResource would be a good project.

> We have an NSF proposal for performance and optimization submitted at  
> the moment, which if funded will definitely include a lot of  
> development of the timer functionality, if it isn't already done.
Wow. Money.

What I'm doing is an immediate fix to make my present job easier, with a
view to other users' experiences with timers.  I see other issues with the
interface and the underlying code.  These could be easily fixed, but we
must do one thing at a time.

Steve White : Programmer
Max-Planck-Institut für Gravitationsphysik      Albert-Einstein-Institut
Am Mühlenberg 1, D-14476 Golm, Germany                  +49-331-567-7329

More information about the Developers mailing list