> Christoph Lameter started by stating that each kernel upgrade resulted in slowdowns for his target applications
You know, Phoronix has access to a wide variety of hardware, does all sorts of tests regularly, and has occasionally done really detailed bisecting work. They also have this "openbenchmarking.org" thing that could almost be a public graph of performance over time for open source projects. It's a real shame the opportunity for something beautiful with regards to detecting performance regressions has been squandered by incoherent graphs, bad statistics, and cheeseball attempts at yellow journalism.