timestamps in fio output for multiple serialized jobs are all showing same time

[bengland2]

For the job file f.job.txt , I get the output f.log. Why are all the timestamps in the output for each process the same? Should the timestamp represent when the fio job actually started running (i.e. after startdelay=N seconds)? If not, what does it mean really? It would be great if the timestamp per job was for the time that the job started issuing I/O.

background: a new Kubernetes-based benchmark called ripsaw is using fio to performance-test Kubernetes pod I/O, and this is basically how they use fio. I just boiled the example down to the simpler form.

[aakarshg]

while I understand that this is to reduce the number of time calls, having timestamps directly in the log files can be desired in certain cases. Especially say in the case where gtod_cpu is enabled.