> + /* We need the for loop, since we need to convert the integer to a float */
> + for (i = 0; i < count && i < desc.RegisterCount; i++)
> + {
> + row[0] = (float)n[i];
The cast is unneeded, and the comment above is too.