You can now get up to 60% higher query throughput (as measured by standard benchmarks TPC-DS, 3TB) in Amazon Redshift as a result of improved memory allocation, which reduces the number of queries spilled to disk. This new improvement is available in version 1.0.1056 and above. Combined with the I/O and commit logic enhancement released in version 1.0.1012, it delivers up to 2 times faster performance for complex queries that spill to disk, and queries like SELECT INTO TEMP TABLE that create temporary tables.