On Thu, Jul 14, 2011 at 02:42:13PM +0200, Francois Tigeot wrote: > > During the course of this run, I noticed system cpu time% was going > through the roof. > I then thought it would be a good idea to re-run the initial benchmarks > with different CPU speeds.
A little bit more about writes: - they were CPU-bound (almost 100% system time used for all CPU threads) - the less CPU power there is in the system, the faster they are -- Francois Tigeot
