Procees Timings

[how...@howard-redway.co.uk]

By default the timings for the top ten (longest) processes are show in the log.  This output can be suppresses by setting the timings parameter to False.  But it is not possible to output the timings for all processes or the top N, where N<>10.

Replacing timings with a parameter to specify the number of processes for which timings are required, with option to show all or none, would be useful.

This would be of value in the following circumstances:

1. Tuning large and complex time critical models
   
2. Advising users of the time penalty of optional proceedures
3. Estimating likely runtimes of proposed model enhancement or whole new models
4. Specifying hardware to achieve required runtimes
5. Increasing the understanding of the factors that determine runtimes.

[Gaëtan de Menten]

On 19/12/2013 12:58, how...@howard-redway.co.uk wrote:
> By default the timings for the top ten (longest) processes are show in

> the log. This output can be suppresses by setting the*/timings
> /*parameter to*/False/*.

Well, no. That option suppresses the output of timings during the
simulation, not the "top ten" at the end.

> But it is not possible to output the timings
> for all processes or the top N, where N<>10.

> Replacing */timings/* with a parameter to specify the number of

> processes for which timings are required, with option to show all or
> none, would be useful.
> This would be of value in the following circumstances:
>

> 1. Tuning large and complex time critical models
> 2. Advising users of the time penalty of optional proceedures
> 3. Estimating likely runtimes of proposed model enhancement or whole
> new models
> 4. Specifying hardware to achieve required runtimes
> 5. Increasing the understanding of the factors that determine runtimes.

Such an option would be really easy to add. It could be a nice
opportunity for you to familiarize yourself with the code. If you have
any trouble to implement that option (or another feature), do not
hesitate to ask for help!

Ga�tan



Disclaimer: please see "www.plan.be/disclaimer.html"

Please consider your environmental responsibility before printing this email