BACKGROUND:
I have been working with a new organisation for the last six weeks and am about to introduce a Win 7 new build.
This will include a newly captured WIM containing pre-reqs, updates etc., a completely new deployment task sequence (SCCM replacing MDT). This is specifically targeting issues of the current build of poor performance and poor reliability.
As the authority on SCCM, the expectations of me are obviously to resolve the above issues - the latter is easy to identify improvements through calls to the desk, but performance (or most importantlyperceived performance) is more subjective.
Therefore, I intend to benchmark their existing build and the new one as a comparison which will hopefully document the benefits in the black and white.
DISCUSSION:
I am interested whether anyone else benchmarks new/amended builds as a matter of course? If so how do you do this?
Traditional benchmarks such as CPU, GPU, RAM are not what I would consider perceived performance indicators. They also vary from model and so are not really relevant to benchmarking the build.
Boot up times and logon times are definitely perceived performance indicators. I will look at xperf (which i see has now been superseded by the Windows Performance Recorder whcih is part of the ADK). I am also aware of the Diagnostics-Performance\Boot Performance Monitoring in the event log.
Can anyone recommend any other benchmark tests that I should include?
Thanks for reading!