[cgl_discussion] Re: [Timers evaluation/Validation] Notes from Timers eval team meetin
rddunlap at osdl.org
Wed Jan 22 16:59:06 PST 2003
On Wed, 22 Jan 2003, Fleischer, Julie N wrote:
| Randy (cc CGL) -
| Here's a summary of what we discussed today re: validation subgroup's role
| in the evaluation of Timers implementations.
| The overall picture is that Geoff will define the background, the POSIX
| requirements, and why we need high-res-timers, Randy is writing the paper
| with the analysis and is performing the engineering evaluation, and Julie is
| doing the validation of current implementations.
| Validation Piece
| At this point, it looks like everything being worked on for POSIX Test Suite
| directly maps to the CGL validation of Timers. A conformance test suite has
| been created (and run against recent versions of both Timers
| implementations). The next test suites to be created are:
| - Functional - Verify full behavior of clocks and timers.
| - Stress/Performance - Verify/Benchmark the behavior of clocks/timers under
| various load conditions.
| We hope to have these tests completed by mid-February, and then a test pass
| can be run against the current implementations. (This should fit within the
| proposed timeline for the project of 1-2 months.)
We would also like to do some kernel profiling of the high-res-timers
code, so if anyone has or knows of one or more programs (or process groups etc.)
that can be used for timer stress testing and for kernel profiling,
please let us know.
More information about the cgl_discussion