Sequential I/O is supposed to wrap-around at the end of the file. The code for that throw a syntax error, and it had some other issues too. I have fixes for all the simple problems already.
It would also be nice if the sequential threads all started in different places too. One approach is to space the processes out evenly, so that if you had 10 threads they'd start at 10%, 20%, etc. This I didn't do yet as I want to look at how some other benchmark utilities handle this first. I did refactor the sequential code already so this sort of change is easier to make next.
Sequential I/O is supposed to wrap-around at the end of the file. The code for that throw a syntax error, and it had some other issues too. I have fixes for all the simple problems already.
It would also be nice if the sequential threads all started in different places too. One approach is to space the processes out evenly, so that if you had 10 threads they'd start at 10%, 20%, etc. This I didn't do yet as I want to look at how some other benchmark utilities handle this first. I did refactor the sequential code already so this sort of change is easier to make next.