> First, it is not clear to me whether a standard should specify this
> kind of functionality. Normally, I would think not, but since
> performance is presumably such an integral part of real-time programs, I
> imagine you could argue that any real-time program will have calls to
> similar functionality in its code, so therefore, standardizing the
> interface will improve program portability. I will leave this issue to
> be debated later.
Real-time performance monitoring will be crucial to the admission test and
quality of service proposal MSU has submitted. The negotiation of channels with
guaranteed quality of service requires a characterization of the traffic the
channel will be generating. These traffic characterization parameters include
average packet transmission interval, average and maximum packet size, etc.
These parameters cannot be entirely known either by the real-time client
software, or the MPI/RT implementation. Instead, the user will have to generate
test runs in MPI/RT, collect traffic information from the RT instrumentation,
and optimize the traffic parameters in the channel admission request based on
this data.
Robert
-- Robert George Engineering Research Center (601) 325-4585 P.O. Box 6176 http://www.erc.msstate.edu/~robert Mississippi State, MS 39762 "Skill and confidence are an unconquered army." - George Herbert