Hi Melba,
> One of the things we need from a charting package is to be able
> to display real-time data that is coming in at 10Hz. One of the
> displays will be a strip chart, which is an x-y plot and all the
> points are plotted and it is moving as new data comes in.
>
> I tried to implement one using one of the tips I found in the
> mail archive about initially creating a large range-sample (10000) with
> NaNs, then as data comes in call setSample(index, data).
> setRange() is set to 500 points and as more data comes in
> setRange() is adjusted by 1.
>
>
> This certainly does what we want but as the data is accumulating
> (around 600 points) cpu usage jumps from 50% to ~99%.
>
> I'm sure if I saved off "old" data the display wouldn't use
> so much cpu to redraw, but our users will want to scroll back...
>
> Is this expected behavior since I'm keeping all the data?
> Do you have any recommendations?
You can cut way down on CPU usage by only updating the Data
values once every N-th point. If that was every 10-th point,
you'd cut CPU use by 90% but still give users an update every
second (at a 10 Hz data rate).
Please let us know if this does not solve your problem.
Good luck,
Bill
