No calculation at runtime is going to speed this query up if you are performing the calculation and then need to filter against the result of the calculation - SQL Server will be forced to perform a table scan. The main problem is that you've chosen to store your dates as a string. For a variety of reasons, this is a terrible decision. Is the string column indexed at least? If so, then this may help get the data only from the last 30 days:
DECLARE @ThirtyDays CHAR(8);
SET @ThirtyDays = CONVERT(CHAR(8),DATEADD(DAY,DATEDIFF(DAY,0,GETDATE()),0)-30,112);
SELECT ...
WHERE cdts >= @ThirtyDays;
If you need to return all the data from all of history except the past 30 days, this isn't going to help either, because unless you are only pulling data from the indexed column, the most efficient approach for retrieving most of the data in the table is to use a clustered index scan. (If you are retrieving a narrow set of columns, it may opt for an index scan, if you have a covering index.) So, your bottleneck in much of these scenarios is not something a formula can fix, but rather the time it takes to actually retrieve a large volume of data, transmit it over the network, and render it on the client.
Also, as an aside, you can't do this:
SELECT a + b AS c FROM dbo.somewhere
WHERE c > 10;
c doesn't exist in dbo.somewhere, it is an expression derived in the SELECT list. The SELECT list is parsed second last (right before ORDER BY), so you can't reference something in the WHERE clause that doesn't exist yet. Typical workarounds are to repeat the expression or use a subquery / CTE.
