For a result set this large, you should make sure you are using remote and not local report processing. Also, if the data source is a normalized, operational database, you should consider moving to a data mart or warehouse and/or an OLAP cube. This is a classic and natural progression. You should also consider implementing scaling options for the database and report servers. You will first need to identify the performance bottlenecks and address them (i.e. memory, processors, index design & usage, 32 vs 64 bit platform, etc.)
Some quick considerations and questions:
Does the query time-out on the database server or only on the client when rendering the report?
What resources are stressed during execution (such as RAM, CPU, cache, etc.)?
Mentor | SQL Server MVP, SolidQ
Author: SQL Server Reporting Services Recipes for Designing Expert Reports; Professional SQL Server Reporting Services (2000, 2005, 2008 & 2012), Beginning T-SQL (2000, 2005 & 2008), SSIS, SSAS & Access