AMD presents "The Bulldozer Blog"

[terrace215]

As for performance, the 50% gain is an aggregate estimation of major server workloads. Estimates tend to be conservative, there is little to be gained from being overly aggressive.

Thank you, that makes the claim have meaning. Although, do you consider HPC-related work to be a "major server workload" ? I ask because that will have the biggest potential increase (~100% core improvement from 128-->256 wide), and so it obviously impacts the mean to the extent it enters into the aggregate.