Indeed, setting .Value on every cell will be slower than it was before, not a whole lot, but if you have a very big load of rows, it can have a hit on performance. This is due to the new mechanics implemented to support cell virtualization, and shutting it off will not change this.
However, this would not happen on a bound grid, that is, the mechanics to fill the grid is a bit different when bound to a data source, which means the grid will perform well, event better than previous versions, when virtualization is turned on.
So the question is, why do you need an unbound grid? Is it absolutely necessary? Usually, an unbound grid is not intended for large DataRow sets, and in your case, it must be quite large since you see the difference in speed.
A solution would be to create an in-between source and bind the grid to it. For example, an ArrayList or whatever, populate it with data from your DataTable, and bind the grid to this ArrayList.
Then you should see the performance gain.
Imported from legacy forums. Posted by André (had 3650 views)