Logging deqp/piglit-runner activity

I've seen this issue several times now: some tests failing when ran through piglit-runner but that pass when run manually or when run using "-t the-failing-test".

I'm wondering if someone else noticed this?

One possible explanation would be that some tests interact in a bad way so I'm wondering if adding a log containing the start/end timestamps of each test (for piglit-runner) and each caselist (for deqp-runner) could be helpful to be able to represent the activity and analyze these failures further.

The output format could be https://docs.google.com/document/d/1CvAClvFfyA5R-PhYUmn5OOQtYMH4h6I0nSsKchNAySU to allow an easy graphical view of the log.