We are experiencing serious performance issues using the RadChart as the number of series increases. At times we receive a Stack Overflow exception from with the charting code. The problem appears to be related to using a Categorical X-Axis. For a Categorical X-Axis it appears that the chart loading time increases exponentially as the # of series increases. Using a Numerical X-Axis does not cause a problem and the load time increases linearly as the # of series is increases.
I created a simple example application that charts a bar series. With only 3 Points per series the load times were as follows:
100 Series (300 total points): 3.4 sec
200 Series (600 total points): 13.9 sec
400 Series (1200 total points): 57.5 sec
800 Series (2400 total points): Stack Overflow exception
As the series count doubles the load time increases 4X. As a comparison, 1200 points in 400 series takes 57.5 sec, but the same number of points in only 1 series (1200 points) plot < 1 sec.
Using a numerical X-Axis the 400 Series x 3 points plots in 10 sec.
I have turned animations off, and data sampling will not help since the problem occurs with very few points per series.
Looking at the call stack during the exception the problem appears to be causes during processing of the categories for the X-Axis. For the above examples the were only a total of 3 categories on the X-Axis.
I am also submitting a support ticket with a sample application.
I created a simple example application that charts a bar series. With only 3 Points per series the load times were as follows:
100 Series (300 total points): 3.4 sec
200 Series (600 total points): 13.9 sec
400 Series (1200 total points): 57.5 sec
800 Series (2400 total points): Stack Overflow exception
As the series count doubles the load time increases 4X. As a comparison, 1200 points in 400 series takes 57.5 sec, but the same number of points in only 1 series (1200 points) plot < 1 sec.
Using a numerical X-Axis the 400 Series x 3 points plots in 10 sec.
I have turned animations off, and data sampling will not help since the problem occurs with very few points per series.
Looking at the call stack during the exception the problem appears to be causes during processing of the categories for the X-Axis. For the above examples the were only a total of 3 categories on the X-Axis.
I am also submitting a support ticket with a sample application.