Should not take that much longer than the C version actually if any. One thing you can try to do to speed things up on the LabVIEW side (if you have not already done so): When you right click the call library function node and go to "Configure..." on the right side is a pull down to specify if or not the DDLs linked function is threadsafe (reentrant) or not. By default the function is always, I believe, assumed to be non-threadsafe which means it will run in the user interface thread which could be a potential slow down point...
I would only do this if you know that the function is threadsafe, however. If it is it should be configured as such, as this leads to performance improvements.