On Thu, Jul 1, 2010 at 11:34 AM, Tres Seaver <tseaver@palladion.com> wrote:
The fact that the test failed seems to make it less than sufficient. I don't quite see how checking pystones is a useful measure, unless you *know* that there is only CPU involved and that the machine where you are checking it is effectively unloaded: I/O and scheduling latencies aren't going to be magically subtracted.
Is this a complaint about the test/documentation or the feature? If the test/documentation, I've already stated what should be done about it -- and I may do so at some point in the near future. In the meantime, adding another zero or five would decrease the chance of it failing again. If the feature: it's certainly not perfect, but has been useful. -- Benji York