At 03:08 PM 2/10/2005, christopher diggins wrote:
The thing is that boost::timer is not good enough for many non-trivial profiling tasks (and it was discussed here on ML many times). Overflow is one of the major issues for example. So without larger time type you couldn't get around it's deficiencies.
It is not a bad idea then to update the timer resolution. However, I find it strange that the expectations on the Profiler library would be greater than those placed on the Boost.Timer library.
Jeff Garland and I have discussed replacing the Boost.Timer library by adding a new timer component in the date-time library. The Boost.Timer design isn't how we would tackle the problem today. There is or was a wiki-page discussing some of the issues, but I'm not finding it at the moment. --Beman